International Journal of Accounting Information Systems 55 (2024) 100709

Contents lists available at ScienceDirect

International Journal of Accounting

Information Systems
journal homepage: www.elsevier.com/locate/accinf

Exploring accounting and AI using topic modelling

Brid Murphy a , Orla Feeney a,* , Pierangelo Rosati b , Theo Lynn c
a
Dublin City University, Dublin 9, Ireland
b
J.E. Cairnes School of Business and Economics, University of Galway, Ireland
c
Irish Institute of Digital Business, Dublin City University, Ireland

A R T I C L E I N F O A B S T R A C T

Keywords: Historically, literature suggests that a variety of accounting roles will be replaced by Artificial
Accounting Intelligence (AI) and related technologies; however, in recent years there is a growing recognition
Accountant that accounting can in fact harness AI’s potential to add value to organisations. Commentators
Artificial intelligence
have highlighted the need for increased research exploring accounting and AI and for accounting
Topic modelling
Latent Dirichlet Allocation
scholars to consider multi-disciplinary research in this area. This study uses a form of topic
modelling to analyse literature exploring AI and related techniques in an accounting context.
Latent Dirichlet Allocation (LDA) has been used to enable probabilistic, machine-based interro­
gation of large volumes of literature. This study applies LDA to the abstracts of 930 peer-reviewed
academic publications from a variety of disciplines to identify the most significant accounting and
AI topics discussed in the literature during the period 1990 to 2023. Our findings suggest that
prior literature reviews based on more traditional methodologies do not capture a comprehensive
picture of accounting and AI research. Eleven topic clusters are identified which provide a
comprehensive topology of the extant literature discussing accounting and AI and set out an
agenda for future research designed to foster academic progress in the area. It also represents one
of the first applications of probabilistic topic modelling to accounting literature.

1. Introduction

For over seventy years, industry and academia have attempted to construct systems that exhibit artificial intelligence (AI) −
systems that think and act like humans (Russell and Norvig, 2010). AI can refer to many things, from analytical techniques (including
machine learning (ML) and deep learning (DL)) to intelligent systems, all exhibiting varying degrees of ‘intelligence’ (Lynn et al.,
2019). Recent commentators have suggested that instead of focussing on a particular definition of intelligence, it may be more useful to
explore the competencies and behaviours demonstrated by AI systems (Kotseruba and Tsotsos, 2016). Others have emphasised the
need to distinguish between artificial general intelligence (AGI) and artificial narrow intelligence (ANI) (Fjelland, 2020; Jiang et al.,
2022). While AGI may exhibit capabilities similar to a human in that it can think abstractly, adapt to change in the environment and
new situations, and perform a wide range of tasks, ANI can only perform a limited range of predefined tasks (Fjelland, 2020, Jiang
et al., 2022). This distinction is important as AGI and ANI are often conflated and may adversely impact our understanding of the role
and impact of AI in accounting.
Initially prevalent in more process-oriented activities, AI is increasingly applied within the knowledge sector, creating an

* Corresponding author.
E-mail addresses: [email protected] (B. Murphy), [email protected] (O. Feeney), [email protected] (P. Rosati), theo.
[email protected] (T. Lynn).

https://doi.org/10.1016/j.accinf.2024.100709
Received 6 July 2023; Received in revised form 3 May 2024; Accepted 12 September 2024
Available online 21 September 2024
1467-0895/© 2024 The Author(s). Published by Elsevier Inc. This is an open access article under the CC BY license
(http://creativecommons.org/licenses/by/4.0/).
B. Murphy et al. International Journal of Accounting Information Systems 55 (2024) 100709

opportunity for accountants to rethink how they engage with their role in an organisational context (Chui et al., 2016). Long held
normative assertions that accounting roles will ultimately be replaced by AI-related technologies (e.g. Frey and Osborne, 2017) are
giving way to a recognition that the accounting function can in fact embrace the growing potential of AI to add value to organisations
(Issa et al., 2016; Issa and Kogan, 2014) i.e. it is now believed that AI can transform the occupational field of accounting as well as
accounting-related tasks (Bauer and Hofmann, 2018; Diller et al., 2020; Lehner et al., 2019; Marrone and Hazelton, 2019; Moll and
Yigitbasioglu, 2019; Murphy and Feeney, 2023). Notwithstanding AI’s potential in this regard, it is important to note that the ac­
counting profession historically lags business adoption of emerging technologies (Dai and Vasarhelyi, 2016).
AI techniques and the nature and use of ML and DL have been the subject of academic focus in a variety of disciplines, notably
computer science, management information systems (MIS) and finance. In their review of accounting research, Sutton et al. (2018)
suggest that, in contrast to Gray et al. (2014), while AI research in accounting did suffer a lull in the early noughties, it has steadily
increased in recent years. There is now a growing volume of research based on a wide range of accounting-related topics including
audit support systems (Sun, 2019), fraud detection (Bao et al., 2020, Papík and Papíková, 2022; Perols, 2011), business intelligence
(Rikhardsson and Yigitbasioglu, 2018), explainable AI (XAI) (Zhang et al., 2022), and accounting service automation (Bavaresco et al.,
2023), amongst others. Given the renewed interest in, as well as reasonable concerns regarding, how AI might change accounting work
and the role of the accountant (ACCA and EY, 2019; AICPA, 2020), it is worthwhile to revisit the literature to evaluate the extant
research on AI in accounting and, based on this review, highlight research gaps and opportunities for future research.
This paper presents an analysis of extant literature exploring AI and related techniques in an accounting context from January 1990
to November 2023 using a topic modelling approach. Topic modelling for structuring literature is a relatively novel innovation in
scholarly research, however it is growing in use in finance, accounting and related fields (Aziz et al., 2022; Dyer et al., 2017; Garanina
et al., 2022; Moro et al., 2015). We follow the approach adopted by Aziz et al. (2022), using the Elsevier Scopus database as the corpus,
and LDA as the topic modelling technique. As such, our study makes two primary contributions. Firstly, it is the first attempt to
structure the topology of the AI and accounting literature and secondly, it is one of the first applications of probabilistic topic modelling
to the accounting literature.
The remainder of this paper is structured as follows. Section 2 defines AI, specifically in an accounting context. Section 3 examines
prior literature reviews on accounting and AI. Our study contributes to this literature by providing an up-to-date typology of research
in accounting and AI, based on an analysis of 930 articles, which identifies emerging trends and an agenda for future research. Section
4 describes the methodological approach and in particular topic modelling and LDA. Section 5 presents the results which set out the
key areas addressed within each topic. The paper concludes with a discussion of potential future avenues for accounting research on AI
as well as the benefits and limitations of LDA as a literature review technique.

2. Defining AI

AI technologies leverage computers and machines to mimic the problem-solving capabilities of the human mind. AI systemises
activities traditionally associated with human intelligence such as planning, learning, reasoning, problem solving, knowledge rep­
resentation, perception, manipulation, even social intelligence, and creativity (Autor and Dorn, 2013; Autor et al., 2003; Frey and
Osborne, 2017). The key difference between an AI and a non-AI application is that AI tools learn to do their job and advance based on
experience without the need to be explicitly programmed (Bolton et al., 2018). Russell and Norvig (2010) distill the various definitions
of AI into two distinctions i.e., human vs rational and think vs act, but their typology has faced criticism for being too simplistic and
imprecise. For example, Wang (2008) suggests that AI should be categorised in five ways – structure, behaviour, capability, function
and principle. In contrast, Kaplan and Haenlein (2019) categorise AI systems by types of competencies i.e., cognitive intelligence,
emotional intelligence, social intelligence, and artistic creativity.
More recently, there has been an increased focus on differentiating between different types of AI. As discussed, ANI (Artificial
Narrow Intelligence) can only perform a limited range of defined tasks and is sometimes referred to as ‘weak AI’ (Fjelland, 2020; Jiang
et al., 2022). ANI represents most AI applications in use today. AGI (Artificial General Intelligence) or ‘strong AI’ is AI that exhibits or
mimics the intelligence capabilities of a human, meaning it can reason, plan, learn from experience, and solve problems autonomously
for tasks that it was never even designed for (Kaplan and Haenlein, 2019). In turn, ASI (Artificial Super Intelligence) is an AI
significantly more intelligent than humans in all respects (Barrett and Baum, 2017).
AI is not an entirely new phenomenon in an accounting context. First generation AI in the form of expert systems, knowledge-based
systems and intelligent systems has existed for decades. Expert systems developed in the 1980s attempted to replicate human expertise
and transform it into rules to perform accounting tasks (Gregor and Benbasat, 1999). It is fair to say that they didn’t live up to their
potential (O’Leary, 1991), probably because they were based on ‘if-then’ rules and decision-trees which frequently codified flawed
logic facilitating the same mistakes to be made over and over (Makridakis, 2017). However, with AI supporting these knowledge-based
systems, together with more emphasis on data analytics and the associated use of ML techniques, increased use of AI in accounting is
not only more likely, but is inevitable (Sutton et al., 2016). As AI-based systems learn to do their jobs and advance based on experience,
much like a human professional, their use will progress from supporting repetitive, routine accounting tasks (Cooper et al., 2019) to
more complex and novel situations requiring flexibility and reasoning (Autor and Dorn, 2013; Brynjolfsson and McAfee, 2014; Frey
and Osborne, 2017; Skrbiš and Laughland-Booÿ, 2019).
What is still unknown at this stage is exactly what human work in accounting will be transferred to AI software and services. Prior
research focused on accounting and the automation of data processing and transaction-based activities (Ghasemi et al., 2011; Tai­
paleenmäki and Ikäheimo, 2013). However, the digital transformation of the accounting function using AI and Big Data analytics
increases the network of people and software, both within and outside the organisation, creating more data pools and decision points

2
B. Murphy et al. International Journal of Accounting Information Systems 55 (2024) 100709

than ever before (Marrone and Hazelton, 2019). Arguably, the real value of AI and Big Data analytics generally, and AI specifically, will
be in enhanced planning, control, forecasting, and decision-making using exponentially larger data pools within a significantly broader
information ecosystem (Brynjolfsson et al., 2011; Feeney, 2022; Vial, 2019). Sutton et al.’s (2016, p.70) suggestion that accounting
researchers ‘have a responsibility to step back and consider the ramifications (of AI) for the future of accounting professionals’ led to
several studies which sought to understand the changing nature of roles in accounting in light of the digital transformation of society
(Leitner-Hanetseder et al., 2021; Moll and Yigitbasioglu, 2019; Oesterreich and Teuteberg, 2019). Our paper contributes to this
literature stream.

3. Prior literature reviews in accounting and AI

Several authors have previously examined literature on accounting and AI, using a variety of approaches. In so doing, they provide
some understanding of the evolution of this topic, its theoretical underpinnings, and some insights into potential new directions.
Mardini and Alkurdi (2021) employed the qualitative panel dimensions approach, which comprised detailed discussion of prior studies
through panels and narrative review, with each panel listing most of the critical prior studies that discuss the uses of AI in accounting
fields. Mardini and Alkurdi (2021) structured the literature into five fields – AI and financial accounting, AI and auditing, AI and cost
and managerial accounting, AI and tax accounting, and AI and public accounting. Hasan (2021) adopted a semi-systematic or narrative
review approach to examine how AI fits into the accounting and audit profession/discipline, as well as benefits and risks associated
with AI implementation for the period 1992–2020. Hasan (2021) identified 10 focus areas – expert systems, continuous auditing,
decision support systems, neural networks, DL and ML, natural language processing (NLP), fuzzy logic, genetic algorithms, robotic
process automation, and hybrid systems. A significant weakness in these focus areas is the overlap between them e.g. one might
reasonably argue that a number of these areas are sub-categories of ML and DL e.g. neural networks, NLP etc. Berdiyeva et al. (2021)
provided a qualitative review of research papers published between the years 1989–2020. While this review is largely descriptive, it
attempts to classify whether extant research suggests a positive, negative, or neutral impact of AI in the accounting and finance process;
they find that the research suggests a strongly positive impact. Ranta et al. (2023) adopted a qualitative approach to provide a
comprehensive synopsis of different ML techniques that might be used to conduct research in management accounting. They note that
there has been progress in the accounting literature in relation to ML and AI in three fields: (1) how AI will change the field of ac­
counting and the development of the accounting profession, (2) textual analysis related to accounting data/reports, and (3) prediction
methods.
Other authors have completed literature reviews on specific sub-themes of accounting and AI literature. For example, Kroon et al.
(2021) undertook a systematic literature review on the impact of emerging technologies (including AI, Big Data analytics, Blockchain
etc.) on accountants’ role and skills, using nVivo to analyse 40 articles published between 2015 and 2020. They adopted the Preferred
Reporting Items for Systematic review and Meta-Analysis (PRISMA) Statement approach. Lehner et al. (2022) conducted a
theoretically-informed, narrative, semi-systematic literature review spanning 2015–2020, using ATLAS.TI qualitative coding software,
and based on Rest’s (1986) four-component model of antecedents for ethical decision-making. Lehner et al. (2022) identified five
categories of ethical challenges – objectivity, privacy and data protection, transparency, accountability, and trustworthiness. Kur­
eljusic and Karger’s (2023) systematic literature review of 47 articles resulted in three key themes centred on AI-based forecasting in
financial accounting – bankruptcy, financial analysis, and frauds and errors. Agustí and Orta-Pérez (2023) generated an overview of
the AI (and blockchain) literature in the fields of accounting and auditing using a bibliometric co-word analysis on 247 research papers
published between the years 1986–2020 using VOSviewer software. They extracted the bibliometric network based on the co-
occurrences of keywords by calculating similarity relations, generating five clusters of words and creating a graphic mapping of the
co-word network. The five clusters emphasise auditing including Big Data analytics and issues related to fraud, risk, and disclosures
(Cluster 1), Big Data and social media, blockchain, or audit evidence (Cluster 2), accounting and AI including applications and ac­
countant training (Cluster 3), ratio prediction and analysis for bankruptcy predictions (Cluster 4), and information technology and
decision-making processes (Cluster 5). Belfiore et al. (2022) analysed a dataset of 3,836 articles published from 2010 to 2021 based on
the co-occurrence of key terms describing the content of documents in a dataset, using the Bibliometrix R package. Their thematic
network is drawn on a two-dimensional matrix, where the axes are functions of relevance (centrality) and density (Callon et al., 1991).
This facilitates the generation of motor themes, basic themes, emerging or declining themes, and niche themes. They identify three
basic themes – quality, impact, and management – and one emerging or declining theme, information. It is important to note that this
review is very high level, thereby limiting utility for future research.
Recent commentators have noted that the use of ML and related methods for accounting research remains in its infancy (Ranta
et al., 2023). Ranta et al. (2023) specifically discuss the use and value of ML techniques for literature reviews and note that LDA has
been used to identify and analyse latent themes in extant literature. However, they do not cite any such reviews related to accounting
and AI. To address this gap and in contrast to extant literature reviews on accounting and AI, this paper uses LDA to structure a to­
pology of the accounting and AI literature and in so doing uncovers latent themes within the research base.

4. Methodology

Topic Modelling is a statistical ‘modelling approach’ which enables the unsupervised discovery of abstract topics that occur in an
unstructured collection or corpus of documents (Blei, 2012). Topic modelling can be non-probabilistic and probabilistic. While non-
probabilistic approaches are still in use (see, for example, Anandkumar et al., 2012; Arora et al., 2013), probabilistic topic modelling
approaches are increasingly the norm as such approaches allow documents and terms belonging to different topics i.e., ‘the same word

3
B. Murphy et al. International Journal of Accounting Information Systems 55 (2024) 100709

might be assigned to different topics in the same document’ (Boyd-Graber et al., 2017, p.14). As such, they provide more nuanced and
realistic results than non-probabilistic approaches. The two main probabilistic topic modelling approaches are LDA (Blei et al., 2003)
and Probabilistic Latent Semantic Analysis (PLSA) (Hofmann, 1999). LDA is more widely used due to easier inference and simpler
model fine-tuning (Boyd-Graber et al., 2017) and has been used in several studies for structuring research literature in finance, ac­
counting and related fields (Aziz et al., 2022; Dyer et al., 2017; Garanina et al., 2022; Moro et al., 2015). In this paper, we apply LDA to
the abstracts of academic publications to identify topics discussed in the literature.
LDA assumes that a document reflects a collection of topics, and the words used by authors are representative of these topics (Blei
et al., 2003; Blei, 2012). In other words, text documents using a similar combination of words relate to the same ‘topic’. In fact, the LDA
algorithm looks for a predefined number of topics in a corpus of documents, determines these topics based on word co-occurrence,
allocates each document to the topic with the highest probability, and provides a list of the most prominent words for each topic.
At this point, the domain knowledge of the researcher is required to interpret the results by working backwards from the list of
observed words to the latent topics.
Documents pre-processing is a multi-step process that is of primary importance for the successful implementation of LDA and to
avoid biased results. To prepare the corpus of documents for the analysis, documents are ‘tokenized’ meaning that each document is
broken down into a collection of each individual word it contains regardless of the order they appear in. In the following step, tokens
that do not represent words (e.g. special characters, punctuation marks, numbers etc.) and so-called ‘stop words’ (i.e., words that
connect different parts of a sentence or that do not carry a meaning per se) are removed (e.g. ‘and’, ‘or’, ‘if’ etc.). The remaining words
are then ‘stemmed’, meaning that suffixes are removed based on the assumption that words with the same stem would recall the same
concept (e.g. “accounting” and “accounts” would both be reduced to “account”); this is an important step as it reduces the word count
(and therefore the complexity of the model) and improves the identification of unique terms within documents (and therefore the
detection of topics). The next step consists of a Term Frequency-Inverse Document Frequency (TF-IDF) analysis which assesses the
importance of a word to a document in a corpus (Salton and Buckley, 1988). This is measured by calculating the percentage of oc­
currences of any term in a given document compared to all other terms in that document and multiplying it by the logarithm of the
number of documents in the corpus divided by the number of documents in the corpus that contain the term. Terms with a higher TF-
IDF measure are more important in the corpus than terms with a lower TF-IDF measure. At this step, terms with very low or very high
TF-IDF are typically removed as they are too uncommon or occur too often in the corpus to describe individual topics. The final step of
this process consists of the construction of a document-term matrix (DTM) where each row represents a document, each column
represents a term in a document, each term and the values reported in the matrix are the frequency of occurrence of a term in a
document. The LDA will then take the DTM as an input and will reduce the number of columns to the predefined number of topics by
grouping together frequently co-occurring terms into a topic. This last phase of the analysis is described in more detail in Blei et al.
(2003) and Boyd-Graber et al. (2017).
In this study, 11 clusters of words were extracted using the LDA model. These clusters were based on text similarity to facilitate the
interpretation and presentation of the results. More specifically, text similarity between clusters was measured using the Hellinger
distance as it has been proven to perform better than other measures in high-dimensionality applications like the one presented in this
study (Sohangir and Wang, 2017). Clusters with higher text similarity (and lower distance) were grouped together to structure the
topography of a corpus of documents spanning multiple disciplines and over 34 years from 1990 to 2023. The source of the corpus was
the Elsevier Scopus database. We used this database to identify studies related to AI or ML and accounting by applying the search
criteria presented in Table 1.
Retrieved articles had to feature any of the AI/ML search terms and any of accounting search terms within the title, abstract, or
indexing keywords. The initial list included 2,877 articles published in 928 different outlets. Due to the frequent use of the word
“accounting” as a verb, a number of non-relevant papers were included in this original list. Two independent coders manually clas­
sified the abstracts as relevant or not (Cohen’s kappa = 0.993); disagreements were reviewed by a third independent coder to agree on
the final classification. The final list included 930 abstracts published in 478 discrete outlets. Fig. 1 provides an overview of the time
distribution of the number of publications in our final list by type (peer reviewed journals and conference proceedings).
When running an LDA algorithm, the number of topics is not known to the researchers a priori. To identify the optimal number of
topics in the corpus of documents, we estimated separate models with the number of topics (k) ranging from 1 to 20 in increments of
one. The optimal number of topics was then selected using the density-based approach presented in Cao et al. (2009) according to
which the best model for a given corpora is the one that maximises the similarity within each topic (measured using cosine similarity)

Table 1
Search criteria.
Search Terms AI/ML search terms: “machine learning” OR “artificial intelligence” OR “support vector machine” OR “deep learning” OR “neural network” OR
“A.I.” OR “AI”
Accounting search terms: “Accounting” OR “Accountant” OR “Auditor” OR “Audit Reporting” OR “Management Reporting” OR “Accounting
Information Systems” OR “Corporate Governance” OR “financial reporting”
Date Range January 1990 to November 2023
Publication Articles and conference papers
Type
Source Type Journals and conference proceedings
Language English language only
Subjects Computer science; business, management and accounting; decision sciences; economics, econometrics and finance.

4
B. Murphy et al. International Journal of Accounting Information Systems 55 (2024) 100709

Fig. 1. Publication Distribution by Year and Publication Type.

while minimising similarity across topics (measured using a perplexity score). The results of this analysis suggest that 11 different
topics could be identified in the corpus of documents. Finally, we ran the LDA algorithm using the ‘topicmodels’ package in R (Grün
and Hornik, 2011). We implemented variational expectation–maximization inference (VEM) (Blei et al., 2003) with 10 different
starting points (seeds) to minimise the Krippendorff’s α (Krippendorff, 1970); lower values of α identify topic models with sharper topic
distributions (Aziz et al., 2022) meaning that while topics may occasionally overlap, they are less likely to do so and therefore easier to
logically separate when interpreting the results. As a result, the final parameters were k = 11 and seed = 10.
While the LDA technique uses quantitative methods to organise the content of a large dataset into topics, a qualitative approach is
needed post-LDA in order to interpret and validate results and align with previous or well-established human-defined categories. This
qualitative phase required the authors to first identify the themes generated by the LDA. This involved scrutinizing the abstracts within
the topics, contextualising them within the broader research context and linking the key themes within the different papers so as to
understand each topic. Two of the authors carried this out independently, taking half of the topics each, first familiarizing themselves

Fig. 2. Distribution of Peer-Reviewed Journal Articles by Discipline and Year.

5
B. Murphy et al. International Journal of Accounting Information Systems 55 (2024) 100709

with the data by reading the abstracts within the topics, labelling the abstracts with potential themes within the topics and grouping
papers into patterns and subcategories within those topics. The two authors then validated each other’s analyses before bringing all the
analyses together in order to set out clear definitions and descriptive names that captured the 11 topics. This final phase involved a
lengthy and iterative process of fine tuning which necessitated a degree of domain expertise on the part of the researchers conducting
the analysis. Topics are described and discussed in the next section.

Table 2
Results Overview.
Topic Label Topic keywords (stemmed) Number of Publication Type Sample Articles – topic match
publications %

1 Neural network forecasting Neural, model, network 102 PRJ − AIS: 1 % Spear and Leis (1997) 100 %
PRJ − Accounting: Lin (2009) 97 %
11 %
PRJ − Other: 85 %
Conferences: 3 %
2 Machine learning and stock Machine learning, risk, stock 91 PRJ – AIS: 1 % Bali et al. (2023) 98 %
prediction PRJ – Accounting: 18 Hanauer and Kalsbach
% (2023) 98 %
PRJ – Other: 68 %
Conferences: 13 %
3 Machine learning and fraud Fraud, financial, data 99 PRJ − AIS: 2 % Moepya et al. (2016) 80 %
detection PRJ − Accounting: Hamal and Senvar (2021) 77
15 % %
PRJ − Other: 66 %
Conferences: 17 %
4 Machine learning in an audit Audit, auditor, auditing 75 PRJ − AIS: 8 % Bertomeu et al. (2021) 89 %
environment PRJ − Accounting: Ragothaman et al. (2000) 76
41 % %
PRJ − Other: 35 %
Conferences: 16 %

5 AI applications in financial Financial, information, 73 PRJ − AIS: 8 % Bos and Frasincar (2022) 82
analysis company PRJ − Accounting: %
26 % Bannier et al. (2019) 81 %
PRJ − Other: 56 %
Conferences: 10 %
6 AI and strategic decision-making Governance, corporate, social 63 PRJ − AIS: 2 % Škapa et al. (2023) 84 %
PRJ − Accounting: 5 Lin (2021) 83 %
%
PRJ − Other: 76 %
Conferences: 17 %

7 AI and the accountant Development, human, 90 PRJ − AIS: 11 % Aldredge et al. (2021) 100 %
professional PRJ − Accounting: Reepu (2020) 95 %
13 %
PRJ − Other: 46 %
Conferences: 30 %
8 AI and accounting systems Financial, information, 110 PRJ − AIS: 0 % Jiang et al. (2022) 95 %
integration enterprise PRJ − Accounting: 0 Jin (2024) 96 %
%
PRJ − Other: 52 %
Conferences: 48 %
9 AI and business operations Business, system, decision 81 PRJ − AIS: 0 % Nado et al. (1996) 94 %
PRJ − Accounting: 4 Tater et al. (2022) 88 %
%
PRJ − Other: 58 %
Conferences: 38 %

10 AI adoption and information Information system, influence, 77 PRJ − AIS: 8 % Khalil and Zainuddin (2015)
quality quality PRJ − Accounting: 98 %
16 % Utomo et al. (2020) 98 %
PRJ − Other: 62 %
Conferences: 14 %
11 Research and education Research, literature, future 69 PRJ − AIS: 25 % De Villiers (2021) 99 %
PRJ − Accounting: Muehlmann et al. (2015) 93
32 % %
PRJ − Other: 39 %
Conferences: 4 %

6
B. Murphy et al. International Journal of Accounting Information Systems 55 (2024) 100709

5. Results

The results present an analysis of papers focusing on AI and accounting which span 34 years (1990–2023) and published in a wide
range of disciplinary journals. It is worth noting that the number of publications has increased significantly in the last four years
covered by our analysis (2020–2023), suggesting a dramatic growth in academic interest in the area of accounting and AI (See Fig. 1
above).
We categorise the peer-reviewed journal articles which emerged from the model across three categories: (1) articles published in
AIS journals (e.g. International Journal of Accounting Information Systems, International Journal of Digital Accounting Research,
Journal of Emerging Technologies in Accounting); (2) articles published in accounting journals (e.g. European Accounting Review,
Journal of Accounting Research; Review of Accounting Studies); (3) articles published in other journals (e.g. Expert Systems with
Applications, Journal of Risk and Financial Management, Sustainability). Fig. 2 presents an overview of the distribution of peer-
reviewed journal articles by journal category and publication year.
Eleven topics pertaining to accounting and AI were extracted. Table 2 provides an overview of the topics, corresponding keywords
and some exemplar papers. It also includes a column titled “Publication Type” that provides a breakdown of the relative frequency of
the number of publications across the three highlighted journal categories.
Topic 1 − Neural Network Forecasting
Topic 1 deals with the use of neural networks for cost estimation, forecasting and prediction. The 102 papers in this topic span the
broadest time frame, with the first published in 1994 and the most recent in late 2023, and include some early seminal papers in the
area. Exemplar keywords include ‘neural’, ‘model’ and ‘network’. The purpose of many of the earlier papers was to introduce the
technology and explain how it can be applied. As a result, the majority of papers appear in engineering, technology and computing
journals and so are categorized in ‘Other Journals’ in Table 2 above. Conventional statistical methods assume linearity, normality and
independence among input variables. These assumptions rarely apply to actual financial data, which limits their validity. ‘Soft’
computing techniques such as probabilistic neural networks and back-propagation neural networks replicate the way the human brain
works. This means that instead of ‘hard’ coding results into a computer using restrictive models or ‘if-then’ coding, an ANN is set-up
with weights, initially random numbers. As it is trained (i.e. engages with actual data) the network learns by adjusting the weights in
positive ways when it achieves a correct outcome, and in negative ways when it achieves an incorrect outcome (Kumar and Bhatta­
charya, 2006). Such learning is particularly useful when a model or ‘if then’ relationship is unknown. Spear and Leis (1997) was one of
the earliest papers to explore how neural network models can be fine-tuned based on the error rates obtained in prior iterations,
leading to more reliable modelling and enhanced capacity for generalisation. In most cases, these articles report on studies which
compare the accuracy of cost prediction over a project’s life using neural networks with actual results or predicted results attained
through non-AI applications, and emphasis is on the technology as much as the accounting application.
Prominent application domains include construction, project and engineering contexts (Juszczyk and Leśniak, 2019; Odeyinka
et al., 2013). Other studies in this topic conduct similar analyses of the efficacy of neural networks in day-to-day cost estimation, e.g.
De Cos et al. (2008) explore the use of an ANN technique for the rapid cost estimation of turbine components eliminating the need for
expert intervention, while Karaoglan and Karademir (2017) examine the predictive capability of neural networks in determining the
production cost of orders with a view to generating more accurate price offers at the tender stage. Several studies also discuss the
application of neural networks in financial analysis in a more general context, e.g. Barney et al. (1999) and Gunawardana (2021) for
debt failure prediction, Lin (2009) and Abrol et al. (2023) for bankruptcy prediction and Florez-Lopez and Ramon-Jeronimo (2009) for
market positioning, while Liang et al. (1992) examine the use of neural networks to make inventory accounting decisions (i.e. LIFO/
FIFO).
Topic 2 – Machine Learning and Stock Prediction
Exemplar keywords in Topic 2 include ‘machine learning’, ‘risk’ and ‘stock’. Papers are focused largely on the potential for neural
networks to make accurate stock market predictions. The topic comprises 91 papers, 72 of which were published since 2020, high­
lighting a proliferation of these types of studies in the last four years of the study. In early AI applications it was difficult to train neural
networks to outperform more traditional linear approaches, so it was limited to use in only a handful of specialised problems. But as the
field advanced, techniques for learning through ‘deep neural networks’ were developed and continue to develop today. Deep Learning
(DL) allows more abstracted learning where there is no known relationship within the input data (Hastie et al., 2009). A more abstract
level of analysis and pattern recognition is required in order to develop a more meaningful and reliable algorithm. These neural
networks are usually based on nonlinear models that can learn complex patterns in high-dimensional data. These models are more
powerful than non-AI, linear-based models because they can capture more complex relationships between variables. They require
more training data, more computational power, and more tuning in order to obtain good results, but they are better suited to real-
world problems and reveal trends that would otherwise be almost impossible to detect. Again, the papers in this topic are, in the
main, technical examinations of potential applications of DL and a large number of these papers are published in ‘Other Journals’ as
categorised in Table 2.
Most of the studies are concerned with predicting stock price movements. In the main, they compare the accuracy of predicted stock
market returns using neural networks with actual returns or predicted returns attained through non-AI applications. Hanauer and
Kalsbach (2023) compare various ML models to predict emerging stock market returns and find that nonlinear models lead to
economically and statistically superior returns over traditional linear models. Bali et al. (2023) drew on more than 12 million ob­
servations over a 25-year period and also report that nonlinear ML models generated statistically and economically sizable profits in
the long-short portfolios of equity options. Page et al. (2023) use ML to identify skilled equity asset managers, Díaz et al. (2023) to
predict gold risk premium performance and Wang and Hao (2023) to forecast oil prices. A key issue addressed in papers in this topic is

7
B. Murphy et al. International Journal of Accounting Information Systems 55 (2024) 100709

risk. ML is used to predict bankruptcy (Aranha and Bolar, 2023), examine interactions between major equity markets and economic
risk factors (Kocaarslan and Soytas, 2023) as well as operationalize the likelihood of CEO deception (Hyde et al., 2024). Vu et al.
(2023) use ML to identify financial distress predictors for Vietnamese listed firms while Hasan et al. (2022) use advanced ML ap­
proaches to examine the relationship between brand capital and stock price crash risk.
Topic 3 – Machine Learning and Fraud Detection
Exemplar keywords associated with Topic 3 include ‘fraud’, ‘financial’ and ‘data’. Again, most of the 99 papers in this topic apply
ML to historical data and compare predictions with actual results (e.g., Berkin et al., 2023; Hamal and Senvar, 2021; Omar et al. 2017;
Papík and Papíková, 2023). The majority of the papers appear in ‘Other Journals’ in Table 2. In the early 2000s, neural networks
became increasingly important tools for data mining, and the papers on this theme have grown significantly from 2010 onwards. Data
mining predicts future outcomes by identifying patterns in clusters of data and building models of what is happening in that data. Liou
(2008) and Moepya et al. (2016) were seminal papers which examined the efficacy of data mining techniques for early fraud detection.
Dbouk and Zaarour’s (2017) paper was designed to provide guidance to regulators and practitioners on how AI technologies can be
used to detect accounting manipulations. It presents the results of a ML application which detects and then predicts earnings
manipulation. Their application was applied to the financial statements of 53 companies over three years, the first year provided the
‘training data’ feeding the model which was applied to years two and three. The ML approach outperformed a manual auditing
approach. Papers in this topic also explore the use of AI in examining textual content from financial statements to detect corporate
fraud. For example, Hajek and Henriques (2017) apply ML to identify fraudulent misrepresentation in the financial comments section
of financial statements. Dong et al. (2018) builds on such work, but instead of relying on data from financial statements uses systemic
functional linguistics (SFL) theory to tap into unstructured data from financial social media platforms to assess the risk of corporate
fraud.
Within ML, a support vector machine (SVM) is a computer algorithm that ‘learns’, using examples, to assign labels to objects. For
example, a SVM can learn to recognize fraudulent credit card activity by examining a large volume of fraudulent and non-fraudulent
credit card activity reports, or to recognize handwritten digits by examining a large collection of scanned images of handwritten digits.
SVMs are particularly useful for numerical prediction, classification and pattern recognition tasks (Noble, 2006). They are amongst the
most robust prediction models and many of the papers in Topic 3 set out to examine their predictive power when compared with other
approaches such as rough set theory (RST), decision trees, and logistic regression (e.g. Trustorff et al., 2011). A further technique,
particle swarm optimization (PSO), can be used to address some of the limitations of SVMs. A PSO model was first presented by
Eberhart and Kennedy (1995) and loosely modelled on group behaviour, such as birds flocking and fish schooling. It optimizes a
problem by iteratively trying to improve a solution in terms of a given measure of quality. PSO is metaheuristic in that it makes no
assumptions about the problem being optimized and can deal with huge volumes of candidate solutions. However, there is no
guarantee that a solution can ever be found. A key aspect of PSO is its flexibility in terms of solving problems that have different
mathematical constraints and is applied as a parameter to SVM models. Chen (2011) examines how AI can be used in bankruptcy
prediction and applies PSO to obtain a suitable parameter setting for a SVM model which can predict financial distress. Chen (2014)
builds on the same study, using PSO in combination with SVMs to establish a model to better forecast financial failure in Taiwanese
listed companies.
Topic 4 − Machine Learning in an Audit Environment
Articles pertaining to Topic 4 focus on how ML can support a variety of audit activities. Of the 75 papers in this topic, 50 were
published since 2020 indicating a real growth in this research context. It is not surprising that the majority of these papers are
published in ‘Accounting Journals’ as illustrated in Table 2. Many of these papers reflect how audit activities are evolving in response
to AI. Exemplar keywords include ‘audit’, ‘auditor’ and ‘auditing’. Early articles explore the initial application of AI-type technology in
an audit sustainability context, such as expert systems and rules-based decision trees (Ragothaman et al., 2000).
In a conventional approach to programming, programmers ‘tell’ the computer what to do, breaking big problems down into many
smaller clearly defined tasks that a computer can easily perform. As noted above (topic 1), humans do not ‘tell’ the computer how to
solve the problem in a neural network. Instead, it learns from observational data, figuring out its own solution to the problem. This
approach has been found to be particularly effective for fraud detection where there are essentially no clues as to where fraud might be,
i.e., a program can’t be instructed specifically as to what to look for or what to do (Bertomeu et al., 2021; Huang et al., 2022; Lokanan,
2019). Sun and Vasarhelyi (2018) introduces the notion of using data analytics combined with DL in an auditing context in order to
identify business insights from massive volumes of text documents. Sun (2019) further illustrates the value of incorporating DL into the
audit environment using text understanding, speech recognition, visual recognition and structured data analysis.
There is also some discussion on the need for enhanced auditor skills (Brazel and Agoglia, 2007), as well as calls for an increased
culture of innovation within audit firms and the profession more broadly (Manita et al., 2020). However, early fears that auditors
would be replaced by ML and AI more broadly are refuted (Chase and Shim, 1991); instead, the evolving nature of audit practices,
thanks to the supporting role of AI, is acknowledged (Alles and Gray, 2020). Auditors deal with large volumes of information and
evidence and AI can greatly aid audit efficiency by enabling the analysis of large populations of data that heretofore could only have
been audited on a sample basis. AI can also play a role in managing audit risk by focusing on areas that require increased attention,
reducing over-auditing, and achieving audit conclusions more quickly (e.g., Davis, 1996). This facilitates more continuous, real-time
auditing with increased forward-looking information (Zhao et al., 2004). Use of ML can also lead to greater audit effectiveness, with
improved audit quality and enhanced accuracy of audit judgments (Rodrigues et al., 2023). Davis (1996) suggests that while more
experienced auditors might use AI models more selectively, these models are also pertinent to training. The use of a multitude of client
examples can supplement experience from real audits to help audit trainees develop the ability to select more relevant information and
make more efficient and effective judgements. Given the potential social impact of accounting/auditing judgements and decisions, a

8
B. Murphy et al. International Journal of Accounting Information Systems 55 (2024) 100709

need for some caution concerning broader ethical and societal impacts is also reported (Sánchez-Medina et al., 2019).
Topic 5 – AI Applications in Financial Analysis
Exemplar keywords associated with Topic 5 include ‘financial’, ‘information’ and ‘company’. All but 14 of the 73 papers in this
corpus were published since 2019. As illustrated in Table 2, just over half are published in ‘Other Journals’, with approximately a
quarter appearing in ‘Accounting Journals’, reflecting the technical nature of a number of the studies. The primary focus is on textual
analysis and papers largely fall into two categories: content analysis of companies’ financial reports (e.g. Ding et al., 2019; Holowczak
et al., 2019; Mushtaq et al., 2022) and sentiment analysis of publicly available company information (e.g. Bannier et al., 2019;
Eachempati and Srivastava, 2022). With reference to the former, AI programs can interrogate financial statements, providing accurate
and valuable insights in a fraction of the time it would take a team of finance professionals to complete. For example, Minhas and
Hussain (2016) use data mining and linguistic analysis on the narrative sections of 10-K forms to establish a means to identify
falsification in financial text while Back et al. (2001) apply neural networks in the form of self-organising maps to analyse text in­
formation in annual reports and compare it with numerical information.
AI applications can also be used to analyse large volumes of text data in publicly available information and determine the sentiment
or overall emotion expressed in the text, offering more accurate, nuanced, and context-aware insights. In several papers, AI is used to
forecast stock prices and influence stock market decisions based on public sentiment as ascertained from news and social media. For
example, Bos and Frasincar (2022) use ML to extract market sentiment from microblogs, Bannier et al. (2019) measure textual
sentiment from business communications and Eachempati and Srivastava (2022) examine the effect of changing market sentiments
associated with news cycles on asset pricing.
Topic 6 – AI and Strategic Decision-Making
Topic 6 focuses on strategic decision-making and comprises 63 papers. The first paper in this topic was published in 2009, with 57
papers published since 2020, 23 in 2023 alone, reflecting a current and future-oriented focus. The significant keywords are ‘corporate’,
‘governance’, and ‘social’. With huge volumes of information flowing around organisations, a variety of advanced AI techniques and
capabilities, including ANN, neuro fuzzy logics and SVM modelling, may help to structure both internally generated and wider industry
financial information and lead to more strategic and better-quality decision making at high/senior levels in organisations. The articles
examine how senior management in large organisations typically operates within a global context, involving a complex ecology of
legislation and regulations, both impacting and impacted by a variety of externalities. A predominant focus is on how AI can lead to
improved management reporting and decision making by bringing a scientific rigour to both quantitative and qualitative information
analysis. Key decisions where AI modelling can support accounting decision making are in the strategic areas of investment (Creamer
and Freund, 2010), tax evasion (Warner et al., 2015) and corporate culture (Chindasombatcharoen et al., 2023). There is a call for
those involved in strategic decision making, both management in organisations and regulators, to take an active role in understanding
and shaping policies and processes concerning AI (Hilb, 2020), particularly with reference to fintech contexts (Burton, 2020).
The vast majority of papers in this topic appear in the ‘Other Journals’ category in Table 2, driven largely by the emergence of
sustainability as a key focus, with studies falling into two groups. The first demonstrates how environmental applications for AI are
broadening as it looks set to be used as a tool within organisations to enhance their economic growth while also improving their
sustainability outcomes (Antoncic, 2020; Giannarakis et al., 2023; Koh et al., 2013). The second recognises that AI can be used to
identify companies’ sustainability efforts, which may not easily be distinguished from financial statements. AI provides investors with
a tool to comprehensively interrogate vast datasets including social media, news and satellite imaging in order to capture sentiment
and provide insights into a company’s true sustainability efforts (Gupta et al., 2021; Lin, 2021; Škapa, et al., 2023). While research
appears to be at an early stage in this arena, AI looks set to complement traditional analyses in order to improve sustainable decision
making and/or drive more sustainable investment (D’Amato et al., 2022; Liu et al., 2023).
Topic 7 – AI and the accountant
Topic 7 comprises 90 articles and focuses on increased AI innovations within the accounting profession and their impact on ac­
counting professionals. All but eight of the 90 articles were published from 2019 onwards, highlighting a very strong current and
future-oriented focus. As illustrated in Table 2, almost half are published in ‘Other Journals’, reflecting the need for accountant roles to
continue to transform and to interact with broader innovations. Exemplar keywords associated with this topic include ‘development’,
‘human’ and ‘professional’.
The accelerated automation of accounting tasks is examined. Early articles discuss technologies such as expert systems (Baldwin-
Morgan, 1995; Meservy et al., 1992) while more recent articles explore AI as a more disruptive technology (Faulconbridge et al., 2023;
Moll and Yigitbasioglu 2019; Moore and Felo, 2022; Reepu, 2020). The continuing digital transformation, increased innovations to
structure systems within organisations, and the potential for further learning innovation and collaboration between humans and
machines are articulated (Chang et al., 2021; Krishna et al., 2022; Leitner-Hanetseder et al., 2021; Marshall and Lambert, 2018;
Petkov, 2020). Some core contexts are specifically examined, including audit (Henage, 2020; Puthukulam et al., 2021; Zemankova,
2019; Zhu, 2021), management accounting (Andreassen, 2020; Marques et al., 2023; Vărzaru, 2022a,b; Zhang et al., 2023b) and Small
and Medium Enterprises (SMEs) (Gonçalves et al., 2022; Nóbrega et al., 2023; Ulrich and Kratt, 2021).
The overriding consensus is that new technologies such as AI will not replace human accountants, but, rather, will provide an
important collaborative opportunity where machines augment accounting professionals’ work (Andreassen, 2020; Goto, 2021). While
AI may replace lower-end and mid-level personnel who were previously engaged in more basic accounting work, studies reveal it will
lead to increased demand for higher-end professionals (Li, 2019; Tiberius and Hirth, 2019). Korhonen et al. (2021) however urge
caution and suggest that when processes are assessed from a distance, non-programmable tasks and expertise may be misinterpreted as
programmable which may result in unsuccessful outcomes, highlighting that humans are critical to future endeavours. In a similar
vein, Li (2019) highlights the importance of making moral judgements and urges caution in relation to addressing the ethical dilemmas

9
B. Murphy et al. International Journal of Accounting Information Systems 55 (2024) 100709

associated with replacing the human element with machines. On a practical level, accounting professionals will need to intensify their
engagement with rapidly developing AI technologies in order to update skills and competencies, keep pace with innovations, remain
relevant and competitive, add value and ultimately facilitate the more strategic transformation of the accounting profession (Aldredge
et al., 2021; Handoko et al., 2019; Holmes and Douglass, 2022; Moll and Yigitbasioglu, 2019; Yang, 2021).
Topic 8 – AI and Accounting Systems Integration
Topic 8 comprises 110 articles, and all but two were published from 2019 onwards, highlighting a very current emphasis. Inter­
estingly, no articles have been published in mainstream Accounting or AIS journals. Table 2 illustrates that a little more than half of the
articles to date are published in ‘Other Journals’ while the remainder are conference presentations. The articles focus largely on the
increasing use of AI to enhance financial information systems within organisations. Exemplar key words include ‘financial’, ‘infor­
mation’ and ‘enterprise’. A number of different AI techniques may be used to improve the accuracy of and enhance the quality of an
individual enterprise’s accounting information system(s). These include neural networks for large data sets, specifically back­
propagation neural networks used as a training process to feed error rates back through a neural network (e.g. Chen, 2022; Cao, 2023;
Li and Liu, 2022; Zeng, 2022b); SVM, used largely for data classification in smaller data sets (e.g. Jiang et al., 2022; Shao et al., 2022;
Shen et al., 2021; Wang et al., 2021; Zhang, 2023; Zhou, 2021); and PSO, to deal with more complex items and their interactions (e.g.
Malhotra, 2023). Cai (2022) acknowledges the challenges AI brings to the workplace but also in terms of education and training to
prepare for the changes associated with accelerating technological advancements.
By incorporating data from a variety of both internal and external sources e.g., accounting software, business plans, ratio analysis,
comparative analysis, AI can help to enhance accounting and financial information systems and outputs (Feng, 2024; Li, 2022a). Such
optimised information can help organisations to manage risks associated with such accounting informatization (Han, 2022) by
identifying and monitoring key risks, including early warning of financial risks (Chen and Zhang, 2022; Hou, 2022; Li, 2022b; Yang,
2022; Zeng, 2022a), fraud risk (Zhang et al., 2021) and broader security risks (Jin, 2024). Other areas of potential application are also
highlighted e.g. environmental accounting (Zhang and Zhu, 2022), accounting for cross-border trading (Zhou, 2023), and accounting
in SMEs (Wu, 2021; Zhao et al., 2022).
Topic 9 – AI and Business Operations
Topic 9 comprises 81 articles, 62 of which were published since 2019. The remaining 19 articles range from 1991 to 2017 i.e.,
publications illustrate a sustained but increasing examination of this topic since the 1990s (with the exception of some years in the
early 2000s). The vast majority of papers appear in the ‘Other Journals’ in Table 2, highlighting the interdisciplinary nature of the
topic. The articles focus on the use of AI tools and techniques to support businesses. Exemplar topic keywords are ‘business’, ‘system’
and ‘decision’. Technology is a key focus: recent articles report how hybrid decision support systems and frameworks can expand
information processing ability beyond that facilitated by single layer modelling tools. These hybrid systems can effectively cater for
more complex problems and more dynamic contexts. They can also incorporate interdependencies among key constituents and thereby
generate more useful information for decision-making. The integration of AI with other technologies is also examined, e.g., blockchain
(Zand et al., 2020; Zhang et al., 2022) and big data (González-Carrasco et al., 2019) which combine the security, immutability and
traceability of data with the possibilities AI offers.
In terms of accounting and reporting, ML may be used to combine fragmented information across companies, sectors and juris­
dictions and in multiple languages to generate an accounting topology or chart-of-account (Jørgensen and Igel, 2021; Lesner et al.,
2020; Munoz et al., 2022; Zhang and Liang, 2023), and to also aid detection of journal entry anomalies (Zupan et al., 2020). More
broadly, AI may be used to support a wide range of salient accounting-related business applications, including accounts payable (Tater
et al., 2022), costing (Lee and Leung, 2012), order management (Khataie et al., 2011), resource management (Hilmola and Gupta,
2015; Hsu and Hsu, 2008), capital structuring (Östermark, 2015), corporate valuation (Yang et al., 2023), inventory classification
(Kaabi, 2022), expense management (Lecue and Wu, 2017), project management (Al-Tabtabai et al., 1997) and process engineering
(Arif et al., 2020). Prominent application domains include construction, project and engineering contexts (Al-Tabtabai et al., 1997;
Baalousha and Çelik, 2011), advertising (Fan and Delage, 2022) insurance (Zuin et al., 2023), banking (González-Carrasco et al.,
2019), and audit (Nado et al., 1996).
Topics 8 and 9 are closely related, each focusing on the use of AI tools and techniques to enhance financial information systems
(Topic 8) and support business decisions (Topic 9). As noted earlier in the paper, occasionally there is some overlap between topics.
Such overlaps can be caused by words or phrases shared amongst different topics, or papers examining AI technologies in diverse
contexts. For example, there also appears to be some overlap between topics 8 and 9 and topic 1 above, though topic 1 is dominated by
older more seminal papers which introduce the technology and its potential while topics 8 and 9 are more current empirical exam­
inations of AI applications.
Topic 10 – AI Adoption and Information Quality
Comprising 77 articles, Topic 10 examines factors influencing AI adoption and the impacts of AI adoption on information quality. It
demonstrates a significant focus in more recent years − all but 10 of the 77 articles were published from 2019 onwards. As illustrated in
Table 2, the majority are published in ‘Other Journals’, indicating that information quality is impacted by a wide variety of factors. The
significant keywords are ‘information system’, ‘influence’ and ‘quality’. Data were predominantly gathered using survey instruments and
quantitatively analysed using regression and structural equation modelling (e.g. Bonsu et al., 2023).
Effective AI adoption is linked to a number of elements in relation to the wider information system, including firm size, organ­
isational culture, regulatory policies, appropriate planning/implementation activities, intrinsic motivations of key personnel, tech­
nology readiness (e.g. existing accounting automation), top management support, user capability, knowledge readiness, perceived
ease of use, perceived usefulness, trust, and system security (Alathamneh, 2020; Al-Dmour et al., 2019; Khalil and Zainuddin, 2015;
Siew et al., 2020; Sumaryati et al., 2020; Ta and Nguyen, 2020; Utomo et al., 2020). Specific contexts are also examined, including

10
B. Murphy et al. International Journal of Accounting Information Systems 55 (2024) 100709

adoption by SMEs (Abdi et al., 2021), banks (Salameh and Lutfi, 2021), auditors (Handoko and Liusman, 2021), public sector entities
(Solikin and Darmawan, 2023), corporate governance functions (Abdou et al., 2021), and accounting and auditing students (Damerji
and Salimi, 2021).
Reported impacts of AI adoption focus largely on improvements in the quality of information and control systems (Alrjoub et al.,
2023; Qasaimeh et al., 2022; Solikin and Darmawan, 2023) Studies reveal greater efficiencies result from replacement of human roles
by ML-based automation (Bavaresco et al., 2023). Impacts are also examined in relation to other technologies and findings suggest that
the impact of AI adoption is greater than that of big data (Bonsu et al., 2023).
Topic 11 – Research and Education
Comprising 69 articles, Topic 11 focuses on literature reviews charting the emergence of AI as a research strand, as well as de­
velopments within educational settings. All but 10 of the 69 articles were published since 2018, illustrating an exponential growth
pattern since the early 2000s (Apostolou et al., 2014; Barrick et al., 2019; Chiu et al., 2019; Kocsis, 2019; Sutton et al., 2016). Almost
one third of these papers are published in ‘Accounting Journals’ and one quarter in ‘AIS Journals’, reflecting the discipline’s
responsiveness to developments in research and education. Exemplar keywords include ‘research’, ‘literature’ and ‘future’. The articles
predominantly appraise prior literature, comprise largely quantitative methodologies, and discuss the transformative force of tech­
nology on accounting research and accounting education.
The evolution of Accounting Information Systems (AIS) from information to knowledge systems, the interdisciplinary nature of
research, the range of research designs and methodologies and metrics such as citations, citation impact and downloads, are described
(Jones and Alam, 2019; Muehlmann et al., 2015; O’Leary, 2010; Vasarhelyi and Greenstein, 2003). In many cases, bibliometric
analysis is used to examine evolving research themes (Agustí and Orta-Pérez, 2023; Aspiranti et al., 2023; Atayah and Alshater, 2021;
Hutchison et al., 2004; Muehlmann et al., 2015; Sun et al., 2022). The growth in unstructured data accelerates the need for new and
more specific tools and techniques to facilitate research on these unstructured data (Ciampi et al., 2021; Ranta et al., 2023), e.g.,
computational linguistics (El-Haj et al., 2019). Such applications have been used to mine documents to obtain insights, make in­
ferences and to create additional methodologies and artefacts to advance knowledge in accounting, auditing and finance. Optimum
results may be achieved when a combination of tools and techniques are employed (Cai et al., 2019; Mahlendorf et al., 2023). Re­
searchers therefore need to continue to update skills and identify potential future research directions (Bertomeu et al., 2021; Gray
et al., 2014; Mancini et al., 2021; Rikhardsson and Yigitbasioglu, 2018; Sutton et al., 2016).
The need to integrate more innovative AI technology for learners in the accounting curriculum within higher education is also
explored as the disconnect between education and practice widens (Coyne et al., 2016; De Villiers, 2021). The manner in which various
technologies have been and/or should be incorporated into accounting education settings, both conceptually and empirically in
various accounting teaching and learning systems, is explored in both a general context (Coyne et al., 2016; Yang and Zhu, 2023) but
also specifically in relation to management accounting (Jayasinghe, 2021; Nielsen, 2022; Samanthi and Gooneratne, 2023) and audit
(Tiron-Tudor and Deliu, 2022).

6. Discussion and conclusion

The paper uses a ML technique, LDA, to structure the topology of the extant literature discussing accounting and AI. There are a
limited number of literature reviews of similar scope or detail on this theme. Our analysis suggests that prior literature reviews, using
more traditional methodologies, do not capture a comprehensive picture of the accounting and AI literature. Our first contribution is
an analysis of 930 articles over a 34-year period from 1990 to 2023 that presents the most holistic view of the accounting and AI
literature to date. These findings contribute to literature exploring the changing role of accounting in light of the digital transformation
of society and provide some focused avenues for future research.
The volume of Accounting and AI research is growing, particularly in the last four years examined (2020–2023) where exponential
growth was revealed. Over the 34-year period of analysis, the two topics drawing the most attention, i.e. with the most published
articles in peer reviewed journals and conference proceedings, are topic 1 ‘Neural Network Forecasting’ and topic 8 ‘AI and Accounting
Systems Integration’. Over time, the trends of publications reveal some interesting insights. In the 1990s and early 2000s, almost half of
articles published focused on topic 1 ‘Neural Network Forecasting’. This first topic continued to dominate in the 2010–2019 period but
an increasing focus on topic 3 ‘Machine Learning and Fraud Detection’ was also observed. The 2020–2023 period revealed an explosion
of articles, representing over 70 % of the total articles published over the full 34-year period of the study. Analysis of this recent period
reveals that the dominant topic was topic 8 ‘AI and Accounting Systems Integration’, followed by topic 2 ‘Machine Learning and Stock
Prediction’, and topic 7 ‘AI and the Accountant’. Despite its earlier dominance, topic 1 ‘Neural Network Forecasting’ generated the least
papers within this recent time period. Neural networks still garner significant academic attention, but are overshadowed somewhat by
newer foci which have emerged in recent years.
The intersection of accounting and AI has opened up new frontiers of research in the areas of financial analysis, auditing and
decision-making. Our results help to identify specific needs for future research in this rapidly evolving landscape. Within Topics 1 and
2, i.e. ‘Neural Network Forecasting’ and ‘Machine Learning and Stock Prediction’ respectively, we have seen a wealth of research
examining the use of advanced algorithms in the areas of cost estimation, financial distress prediction and stock market forecasting.
Future research in this domain would benefit from increasing datasets to include additional geographies and emerging markets (e.g.
Page et al., 2023, Vu et al., 2023). There is also a need to enhance the interpretability and explainability of AI models in these contexts,
which would help to increase trust amongst practitioners, regulatory bodies and wider stakeholders, which is crucial for a more
widespread adoption of AI.
Topics 3 ‘Machine Learning and Fraud Detection’ and 4 ‘Machine Learning and Auditing’ revealed sustained focus over the period of the

11
B. Murphy et al. International Journal of Accounting Information Systems 55 (2024) 100709

study and increasing growth in associated publications from 2019 onwards. Future research is likely to focus on enhancing anomaly
detection algorithms to better identify irregularities in financial data (e.g. Hamal and Senvar, 2021). Additionally, exploring the
integration of a variety of AI and other techniques for analysing data in financial reports will provide deeper insights (e.g. Sun, 2019).
As professionals charged with providing assurance over financial data, auditors will need to adapt by gaining proficiency in data
analytics, programming and ML in order to leverage technological advancements effectively. This is likely to drive a research agenda.
In addition, ethical considerations pertaining to AI-augmented auditing processes will require greater scrutiny from both practitioners
and researchers to ensure responsible implementation and execution (Sánchez-Medina et al., 2019).
In Topic 5 ‘AI Applications in Financial Analysis’ and Topic 6 ‘AI and Strategic Decision-making’, we present empirical evidence of AI
supporting financial analysis and consequent decisions in a variety of accounting contexts. A key strength of AI in this domain is its
ability to analyse the sentiment of large volumes of varied text data. Future efforts will develop more complex models capable of
discerning context, sarcasm and subtle nuance to provide a more comprehensive understanding of the data (e.g., Bannier et al., 2019),
and enable researchers to better interpret such data. In turn, managers will likely make more optimum decisions which will also
provide a catalyst for further research endeavours. Sustainability emerged as a key research area within Topic 6 in the last four years of
the study. Future research is likely to focus further on the use of AI to help businesses align their economic goals with their envi­
ronmental responsibilities (e.g., Giannarakis et al., 2023) and identify more sustainable investment opportunities (e.g., D’Amato et al.,
2022).
Topic 7 ‘AI and the Accountant’ reveals the transformative effect of AI on accounting as a discipline. As AI accelerates the auto­
mation of routine accounting tasks, the evolving role of accountants remains at the forefront of the accounting agenda. Future studies
should explore not only the skills and competencies required of accountants in an increasing AI-driven era but also how professionals in
an accounting context should work together and negotiate going forward (e.g., Faulconbridge et al., 2023). Again, a primary focus
must be on developing trust mechanisms that enhance the transparency and interpretability of AI-driven accounting information.
Topic 8 ‘AI and Accounting Systems Integration’ and Topic 9 ‘AI and Business Operations’ report on the use of AI within accounting
information systems and wider decision-making systems. Studies suggest that evolving AI techniques have more potential than
traditional techniques to support more complex problems and more dynamic contexts (Feng, 2024). Future research is likely to focus
on an expanding array of AI techniques, both individually and in tandem with other technologies e.g. blockchain (Zhang et al., 2022)
and big data (González-Carrasco et al., 2019), to analyse vast datasets.
Topic 10 ‘AI Adoption and Information Quality’ reports that a wide array of factors have influenced AI adoption to date (e.g. Utomo
et al., 2020). As AI becomes more of an imperative for adoption, future research will need to study contexts that have not yet been
subject to such examination e.g. nonprofit organisations, additional sectors and jurisdictions. With regard to reported impacts of AI
adoption, existing research has focused largely on information quality, efficiencies and control systems (Alrjoub et al., 2023; Bavaresco
et al., 2023). Future research might also consider additional variables and impacts e.g. transparency, fairness, integrity. In addition,
research to date has largely focused on quantitative data analysis: future research could incorporate more qualitative methodologies
and generate a deeper but richer understanding of both AI adoption and impacts in varied settings.
Topic 11 ‘Research and Education’ suggests that research scholars must continuously update skills in order to research more
effectively. Cross-disciplinary research will remain crucial to advancing this research agenda and it is hoped that more qualitative and
mixed method approaches will be embraced so as to develop greater depth of insight. This topic also highlights that the future of
accounting education lies in preparing students to navigate a technologically advanced landscape, ensuring that they possess the skills
to utilise AI while upholding ethical and professional standards i.e. AI literacy must also be integrated into various accounting curricula
(De Villiers, 2021). This will present additional fields for research as students navigate unprecedented evolution.
A topology of research on accounting and AI aids researchers and journal editors by providing a tool for identifying emerging trends
and gaps in the field or a specific journal(s) and enhancing research quality through informed and targeted literature reviews. As such,
it facilitates interdisciplinary collaboration, informs editorial policies, and guides the design of educational curricula. This topology is
also valuable for strategic planning in research, aiding in grant applications, and fostering community building among researchers with
interests in specific topics. Moreover, it promotes transparency, reproducibility, and the effective allocation of resources, ultimately
contributing to the advancement and impact of research at the intersection of accounting and AI.
While more comprehensive than extant literature reviews, there are limitations in our study. Firstly, we limited our analysis to
journal and conference publications in the Elsevier Scopus database. Increasing the corpus to include more publication types (e.g.,
books and book chapters), other databases, and a wider number of research fields, may present even greater insights. Secondly, our
analysis presents a comprehensive overview of what has been studied and provides an agenda for future research addressing the 11
topics in our typology, but structuring corpora into different disciplines (e.g., accounting, banking and finance, economics, IS, com­
puter science) or fields of accounting (similar to Mardini and Alkurdi (2021)) may provide greater comparison or insight. Thirdly, all
literature reviews are a snapshot in time. Even in the last four years of the study, there has been an upsurge in interest with respect to
topic modelling, fuelled by the release and widespread adoption and use of generative AI such as ChatGPT. It is likely the significant
increase in the volume of research in the latter years of our review will continue to accelerate.
Our second contribution relates to our use of probabilistic topic modelling and addressing a gap in the general accounting literature
with respect to the use of LDA for scholarly literature reviews. As discussed, LDA has been used for such purpose in related fields
including banking and finance (Aziz et al., 2022; Moro et al., 2015), economics and management (Elmsili and Outtaj, 2018), and
blockchain in accounting (Garanina et al., 2022). However, it has not been applied widely in accounting, and not specifically to ac­
counting and AI. Our experience with LDA is similar to others in that it has substantial advantages over traditional literature review
methodologies including more efficient topic extraction, organisation and categorization, and exploration. However, it is not without
limitations. This technique involves significant interpretation challenges and therefore LDA still requires domain expertise and careful

12
B. Murphy et al. International Journal of Accounting Information Systems 55 (2024) 100709

analysis to interpret and validate results and align with previous or well-established human-defined categories. Similarly, as LDA is
sensitive to the number of topics chosen and other hyperparameters, there is still a degree of manual fine-tuning and cross-validation
for results. Researchers seeking to use LDA should be aware that LDA ignores word order and contextual information, thereby hin­
dering the understanding of relationships and contextual nuances a human or other techniques may identify. A typical challenge with
LDA is the quality of input data. In the case of scholarly publications, the quality and consistency of input data is typically higher than
other input sources and so this was not a significant issue in our study. Notwithstanding, care should still be taken that appropriate
inclusion/exclusion criteria are chosen, and pre-processing undertaken.
Recent advancements in neural topic models based on advanced NLP architectures using transformers (e.g., BERT, GPT and
LLaMA) overcome some of the limitations of classic topic models such as LDA (Churchill and Singh, 2022; Egger and Yu, 2022). These
transformer-based models have greater capacity for contextual understanding, interpretability, topic hierarchy, and modelling large
and very large corpora (Egger and Yu, 2022). However, it is important to note that these models may bring their own challenges, such
as increased computational requirements, the need for large amounts of training data, and the costs of labelling such data (Churchill
and Singh, 2022; Egger and Yu, 2022; Zengul et al., 2023). Furthermore, they may generate large numbers of topics and outliers, which
may ultimately result in greater time and effort (Egger and Yu, 2022). There is increasing evidence that transformer-based models may
also result in hallucinations i.e., output that conflicts with the source or cannot be verified by factual knowledge, which raises concerns
regarding the accuracy of results (Huang et al. 2023; Zhang et al., 2023a). LDA and transformer-based large language models (LLMs)
may be used as part of a wider pipeline. In this scenario, LDA might be used for broader topic modelling and transformer-based LLMs
used for deeper analysis within each topic. Given the novelty of neural topic modelling, there are relatively few applications to
literature reviews in the business research domain. Echoing Eickhoff and Neuss (2017), more implementations and comparative
evaluation studies are required to benchmark and validate these techniques and their suitability in different literature review contexts,
both separately and as part of a wider NLP pipeline.
In conclusion, topic modelling is a useful tool for the efficient analysis of scholarly literature and the identification of latent patterns
and themes. While advancements in natural language modelling and DL promise even greater contextual understanding and analytical
power, such models still require domain expertise for interpretation and new specialised ensemble or hybrid methodologies for
identifying under-researched areas for future research.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to
influence the work reported in this paper.

Data availability

No data was used for the research described in the article.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.accinf.2024.100709.

References

Abdi, M.D., Dobamo, H.A., Bayu, K.B., 2021. Exploring current opportunity and threats of artificial intelligence on small and medium enterprises accounting function;
evidence from southwest part of ethiopia, oromiya, jimma and snnpr, bonga. Acad. Acc. Financ. Stud. J. 25 (2), 1–11.
Abdou, H.A., Ellelly, N.N., Elamer, A.A., Hussainey, K., Yazdifar, H., 2021. Corporate governance and earnings management nexus: evidence from the UK and Egypt
using neural networks. Int. J. Financ. Econ. 26 (4), 6281–6311.
Abrol, V., Singh, P., Subhashini, R., Kumar, A.S., Singh, K., Mannar, B.R., 2023. Bankruptcy prediction using emperor penguin optimizer with deep learning model on
qualitative dataset. In: 2023 5th International Conference on Smart Systems and Inventive Technology (ICSSIT). IEEE, pp. 1028–1033.
ACCA, EY, 2019. The Impact of Digital and Artificial Intelligence on Audit and Finance Professionals: Harnessing the Opportunities of Disruptive Technologies.
Association of Chartered Certified Accountants (ACCA).
Agustí, M.A., Orta-Pérez, M., 2023. Big data and artificial intelligence in the fields of accounting and auditing: a bibliometric analysis. Span. J. Finance Acc. 52 (3),
412–438.
AICPA, 2020. The Data-Driven Audit: How Automation and AI are Changing the Audit and the Role of the Auditor. Association of Chartered Professional Accountants
(AICPA).
Alathamneh, M., 2020. The impact of accounting information systems reliability on enhancing the requirements of planning process at Jordanian commercial banks.
Manage. Sci. Lett. 10 (5), 1043–1050.
Al-Dmour, A.H., Abood, M., Al-Dmour, H.H., 2019. The implementation of SysTrust principles and criteria for assuring reliability of AIS: empirical study. Int. J.
Account. Inf. Manag. 27 (3), 461–491.
Aldredge, M., Rogers, C., Smith, J., 2021. The strategic transformation of accounting into a learned profession. Ind. High. Educ. 35 (2), 83–88.
Alles, M.G., Gray, G.L., 2020. Will the medium become the message? A framework for understanding the coming automation of the audit process. J. Inf. Syst. 34 (2),
109–130.
Alrjoub, A.M.S., Bataineh, A., Al Qudah, L.A.M., Al Othman, L.N., Alkarabsheh, F., Aburisheh, K.E., 2023. The impact of quality costs as a mediator in the relationship
between management accounting systems and financial performance: the case of Jordan. Int. J. Professional Business Rev. 8 (4), 30.
Al-Tabtabai, H., Kartam, N., Flood, I., Alex, A.P., 1997. Construction project control using artificial neural networks. AI EDAM 11 (1), 45–57.

13
B. Murphy et al. International Journal of Accounting Information Systems 55 (2024) 100709

Anandkumar, A., Foster, D.P., Hsu, D.J., Kakade, S.M., Liu, Y.K., 2012. A spectral algorithm for latent dirichlet allocation. Advances in Neural Information Processing
Systems, 25.
Andreassen, R.I., 2020. Digital technology and changing roles: a management accountant’s dream or nightmare? J. Manag. Control. 31 (3), 209–238.
Antoncic, M., 2020. A paradigm shift in the board room: incorporating sustainability into corporate governance and strategic decision-making using big data and
artificial intelligence. J. Risk Manage. Financ. Inst. 13 (4), 290–294.
Apostolou, B., Dorminey, J.W., Hassell, J.M., Rebele, J.E., 2014. A summary and analysis of education research in accounting information systems (AIS). J. Acc. Educ.
32 (2), 99–112.
Aranha, M., Bolar, K., 2023. Efficacies of artificial neural networks ushering improvement in the prediction of extant credit risk models. Cogent Econ. Finance 11 (1),
2210916.
Arif, D., Yucha, N., Setiawan, S., Oktarina, D., Martah, V., 2020. Applications of goods mutation control form in accounting information system: a case study in
Sumber Indah Perkasa Manufacturing, Indonesia. J. Asian Finance Econ. Business 7 (8), 419–424.
Arora, S., Ge, R., Halpern, Y., Mimno, D., Moitra, A., Sontag, D., Wu, Y., Zhu, M., 2013. A practical algorithm for topic modeling with provable guarantees. In:
International Conference on Machine Learning. PMLR, pp. 280–288.
Aspiranti, T., Ali, Q., Parveen, S., Amaliah, I., Jalil, A., Merican, F.M.I., 2023. Bibliometric review of corporate governance of Islamic financial institutions through AI-
based tools. Int. J. Professional Business Rev. 8 (4), 12.
Atayah, O.F., Alshater, M.M., 2021. Audit and tax in the context of emerging technologies: a retrospective analysis, current trends, and future opportunities. Int. J.
Digit. Account. Res. 21.
Autor, D.H., Dorn, D., 2013. The growth of low-skill service jobs and the polarization of the US labor market. Am. Econ. Rev. 103 (5), 1553–1597.
Autor, D.H., Levy, F., Murnane, R.J., 2003. The skill content of recent technological change: an empirical exploration. Q. J. Econ. 118 (4), 1279–1333.
Aziz, S., Dowling, M., Hammami, H., Piepenbrink, A., 2022. Machine learning in finance: a topic modeling approach. Eur. Financ. Manag. 28 (3), 744–770.
Baalousha, Y., Çelik, T., 2011. An integrated web-based data warehouse and artificial neural networks system for unit price analysis with inflation adjustment. J. Civ.
Eng. Manag. 17 (2), 157–167.
Back, B., Toivonen, J., Vanharanta, H., Visa, A., 2001. Comparing numerical data and text information from annual reports using self-organizing maps. Int. J. Account.
Inf. Syst. 2 (4), 249–269.
Baldwin-Morgan, A.A., 1995. Integrating artificial intelligence into the accounting curriculum. Acc. Educ. 4 (3), 217–229.
Bali, T.G., Beckmeyer, H., Moerke, M., Weigert, F., 2023. Option return predictability with machine learning and big data. Rev. Financ. Stud. 36 (9), 3548–3602.
Bannier, C., Pauls, T., Walter, A., 2019. Content analysis of business communication: introducing a German dictionary. J. Bus. Econ. 89, 79–123.
Bao, Y., Ke, B., Li, B., Yu, Y.J., Zhang, J., 2020. Detecting accounting fraud in publicly traded US firms using a machine learning approach. J. Account. Res. 58 (1),
199–235.
Barney, D.K., Graves, O.F., Johnson, J.D., 1999. The farmers home administration and farm debt failure prediction. J. Account. Public Policy 18 (2), 99–139.
Barrett, A.M., Baum, S.D., 2017. A model of pathways to artificial superintelligence catastrophe for risk and decision analysis. J. Exp. Theor. Artif. Intell. 29 (2),
397–414.
Barrick, J.A., Mecham, N.W., Summers, S.L., Wood, D.A., 2019. Ranking accounting journals by topical area and methodology. J. Inf. Syst. 33 (2), 1–22.
Bauer, W., Hofmann, J., 2018. Arbeit, IT und digitalisierung. In: Hofmann, J. (Ed.), Arbeit 4.0–Digitalisierung, IT Und Arbeit. Springer Vieweg, Wiesbaden, pp. 1–16.
Bavaresco, R.S., Nesi, L.C., Barbosa, J.L.V., Antunes, R.S., da Rosa Righi, R., da Costa, C.A., Vanzin, M., Dornelles, D., Gatti, C., Ferreira, M., Silva, E., 2023. Machine
learning-based automation of accounting services: an exploratory case study. Int. J. Account. Inf. Syst. 49, 100618.
Belfiore, A., Gnasso, A., Cuccurullo, C. and Aria, M., 2022. AI and ML in accounting and finance: a bibliometric review. In: JADT 2022 Proceedings of the 16th
International Conference on Statistical Analysis of Textual Data, 1, pp. 95–101.
Berdiyeva, O., Islam, M.U., Saeedi, M., 2021. Artificial intelligence in accounting and finance: meta-analysis. Int. Bus. Rev. 3 (1), 56–79.
Berkin, A., Aerts, W., Van Caneghem, T., 2023. Feasibility analysis of machine learning for performance-related attributional statements. Int. J. Account. Inf. Syst. 48,
100597.
Bertomeu, J., Cheynel, E., Floyd, E., Pan, W., 2021. Using machine learning to detect misstatements. Rev. Acc. Stud. 26, 468–519.
Blei, S., 2012. On symmetric and skew Bessel processes. Stoch. Process. Appl. 122 (9), 3262–3287.
Blei, D.M., Ng, A.Y., Jordan, M.I., 2003. Latent dirichlet allocation. J. Mach. Learn. Res. 3 (Jan), 993–1022.
Bolton, C., Machová, V., Kovacova, M., Valaskova, K., 2018. The power of human–machine collaboration: artificial intelligence, business automation, and the smart
economy. Econ. Manage. Financ. Markets 13 (4), 51–56.
Bonsu, M.O.A., Wang, Y., Guo, Y., 2023. Does fintech lead to better accounting practices? Empirical evidence. Acc. Res. J. 36 (2/3), 129–147.
Bos, T., Frasincar, F., 2022. Automatically building financial sentiment lexicons while accounting for negation. Cogn. Comput. 14 (1), 442–460.
Boyd-Graber, J., Hu, Y., Mimno, D., 2017. Applications of topic models. Found. Trends Inf. Retr. 11 (2–3), 143–296.
Brazel, J.F., Agoglia, C.P., 2007. An examination of auditor planning judgements in a complex accounting information system environment. Contemp. Account. Res.
24 (4), 1059–1083.
Brynjolfsson, E., Hammerbacher, J., Stevens, B., 2011. Competing through data: three experts offer their game plans. McKinsey Quarterly 4, 36–47.
Brynjolfsson, E., McAfee, A., 2014. The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies. WW Norton & Company.
Burton, D., 2020. Digital debt collection and ecologies of consumer overindebtedness. Econ. Geogr. 96 (3), 244–265.
Cai, C., 2022. Training mode of innovative accounting talents in colleges using artificial intelligence. Mobile Inf. Syst.
Cai, C.W., Linnenluecke, M.K., Marrone, M., Singh, A.K., 2019. Machine learning and expert judgement: analyzing emerging topics in accounting and finance research
in the Asia-Pacific. Abacus 55 (4), 709–733.
Callon, M., Courtial, J.P., Laville, F., 1991. Co-word analysis as a tool for describing the network of interactions between basic and technological research: the case of
polymer chemistry. Scientometrics 22, 155–205.
Cao, P., 2023. Research on the impact of artificial intelligence-based e-commerce personalization on traditional accounting methods. Int. J. Intell. Networks 4,
193–201.
Cao, J., Xia, T., Li, J., Zhang, Y., Tang, S., 2009. A density-based method for adaptive LDA model selection. Neurocomputing 72(7-9), 1775–1781.6.
Chang, J., Hsiao, M., Peng, Y., 2021. An exploration on accounting professionals facing the development of AI. In: Proceedings of the 7th International Conference on
Industrial and Business Engineering (pp. 94–103).
Chase, M.D., Shim, J.K., 1991. Artificial intelligence and big six accounting: a survey of the current uses of expert systems in the modern accounting environment.
Comput. Ind. Eng. 21 (1–4), 205–209.
Chen, M.Y., 2011. A hybrid model for business failure prediction-utilization of particle swarm optimization and support vector machines. Neural Network World 21
(2), 129.
Chen, M.Y., 2014. Using a hybrid evolution approach to forecast financial failures for Taiwan-listed companies. Quantitative Finance 14 (6), 1047–1058.
Chen, Y., 2022. Enterprise financial data sharing based on information fusion cloud computing environment. Wirel. Commun. Mob. Comput. 2022, 1–11.
Chen, Y., Zhang, S., 2022. Accounting information disclosure and financial crisis beforehand warning based on the artificial neural network. Wirel. Commun. Mob.
Comput. 1–11.
Chindasombatcharoen, P., Chatjuthamard, P., Jiraporn, P., 2023. Corporate culture, cultural diversification, and independent directors: evidence from earnings
conference calls. J. Behav. Exp. Financ. 37, 100773.
Chiu, V., Liu, Q., Muehlmann, B., Baldwin, A.A., 2019. A bibliometric analysis of accounting information systems journals and their emerging technologies
contributions. Int. J. Account. Inf. Syst. 32, 24–43.
Chui, M., Manyika, J. and Miremadi, M., 2016. Where machines could replace humans-and where they can’t (yet). McKinsey. Available at: https://www.mckinsey.
com/capabilities/mckinsey-digital/our-insights/where-machines-could-replace-humans-and-where-they-cant-yet.
Churchill, R., Singh, L., 2022. The evolution of topic modeling. ACM Comput. Surv. 54 (10s), 1–35.

14
B. Murphy et al. International Journal of Accounting Information Systems 55 (2024) 100709

Ciampi, F., Giannozzi, A., Marzi, G., Altman, E.I., 2021. Rethinking SME default prediction: a systematic literature review and future perspectives. Scientometrics 126,
2141–2188.
Cooper, L.A., Holderness Jr, D.K., Sorensen, T.L., Wood, D.A., 2019. Robotic process automation in public accounting. Account. Horiz. 33 (4), 15–35.
Coyne, J.G., Coyne, E.M., Walker, K.B., 2016. A model to update accounting curricula for emerging technologies. J. Emerging Technol. Account. 13 (1), 161–169.
Creamer, G., Freund, Y., 2010. Using boosting for financial analysis and performance prediction: application to S&P 500 companies, Latin American ADRs and banks.
Comput. Econ. 36, 133–151.
Dai, J., Vasarhelyi, M.A., 2016. Imagineering Audit 4.0. J. Emerg. Technol. Acc. 13(1), pp. 1–15.
D’Amato, V., D’Ecclesia, R. and Levantesi, S., 2022. Firms’ profitability and ESG score: a machine learning approach. Appl. Stochastic Models Business Industry pp.
1–19.
Damerji, H., Salimi, A., 2021. Mediating effect of use perceptions on technology readiness and adoption of artificial intelligence in accounting. Acc. Educ. 30 (2),
107–130.
Davis, J.T., 1996. Experience and auditors’ selection of relevant information for preliminary control risk assessments. Auditing: A J. Practice Theory 15(1), pp. 16–37.
Dbouk, B., Zaarour, I., 2017. Towards a machine learning approach for earnings manipulation detection. Asian J. Business Acc. 10 (2), 215–251.
de Cos, J., Sanchez, F., Ortega, F., Montequin, V., 2008. Rapid cost estimation of metallic components for the aerospace industry. Int. J. Prod. Econ. 112 (1), 470–482.
De Villiers, R., 2021. Seven principles to ensure future-ready accounting graduates - a model for future research and practice. Meditari Accountancy Res. 29 (6),
1354–1380.
Díaz, J.D., Hansen, E., Cabrera, G., 2023. Gold risk premium estimation with machine learning methods. J. Commod. Mark. 31, 100293.
Diller, M., Asen, M., Späth, T., 2020. The effects of personality traits on digital transformation: evidence from German tax consulting. Int. J. Account. Inf. Syst. 37,
100455.
Ding, P., Zhuoqian, L., Yuan, D., 2019. Textual information extraction model of financial reports. In: Proceedings of the 2019 7th International Conference on
Information Technology: IoT and Smart City (pp. 404–408).
Dong, W., Liao, S., Zhang, Z., 2018. Leveraging financial social media data for corporate fraud detection. J. Manag. Inf. Syst. 35 (2), 461–487.
Dyer, T., Lang, M., Stice-Lawrence, L., 2017. The evolution of 10-K textual disclosure: evidence from Latent Dirichlet Allocation. J. Account. Econ. 64 (2–3), 221–245.
Eachempati, P., Srivastava, P.R., 2022. Accounting for investor sentiment in news and disclosures. Qualitative Res. Financ. Markets 14 (1), 53–75.
Eberhart, R., Kennedy, J., 1995. Particle swarm optimization. In: Proceedings of the IEEE International Conference on Neural Networks (vol. 4, pp. 1942–1948). IEEE.
Egger, R., Yu, J., 2022. A topic modeling comparison between lda, nmf, top2vec, and bertopic to demystify twitter posts. Front. Sociol. 7, 886498.
Eickhoff, M., Neuss, N., 2017. Topic modelling methodology: its use in information systems and other managerial disciplines. In: Proceedings of the 25th European
Conference on Information Systems (ECIS), Guimarães, Portugal, June, (pp. 1327–1347).
El-Haj, M., Rayson, P., Walker, M., Young, S., Simaki, V., 2019. In search of meaning: Lessons, resources and next steps for computational analysis of financial
discourse. J. Bus. Financ. Acc. 46 (3–4), 265–306.
Elmsili, B., Outtaj, B., 2018. Artificial neural networks applications in economics and management research: an exploratory literature review. In: 2018 4th
International Conference on Optimization and Applications (ICOA), pp. 1–6.
Fan, R., Delage, E., 2022. Risk-aware bid optimization for online display advertisement. In: Proceedings of the 31st ACM International Conference on Information &
Knowledge Management (pp. 457–467).
Faulconbridge, J., Sarwar, A., Spring, M., 2023. How professionals adapt to artificial intelligence: the role of intertwined boundary work. J. Manag. Stud.
Feeney, O., 2022. Accounting in a social context, In: Alsharari, N.M. (Ed) Accounting and Finance Innovations, Intech Open, pp. 17–31.
Feng, X., 2024. Data driven mixed industry restructuring of state-owned enterprises in the digital marketing strategy era.
Fjelland, R., 2020. Why general artificial intelligence will not be realized. Humanities Soc. Sci. Commun. 7 (1), 1–9.
Florez-Lopez, R., Ramon-Jeronimo, J.M., 2009. Marketing segmentation through machine learning models: an approach based on customer relationship management
and customer profitability accounting. Soc. Sci. Comput. Rev. 27 (1), 96–117.
Frey, C.B., Osborne, M.A., 2017. The future of employment: how susceptible are jobs to computerisation? Technol. Forecast. Soc. Chang. 114, 254–280.
Garanina, T., Ranta, M., Dumay, J., 2022. Blockchain in accounting research: current trends and emerging topics. Account. Audit. Account. J. 35 (7), 1507–1533.
Ghasemi, M., Shafeiepour, V., Aslani, M., Barvayeh, E., 2011. The impact of Information Technology (IT) on modern accounting systems. Proc. Soc. Behav. Sci. 28,
112–116.
Giannarakis, G., Mallidis, I., Sariannidis, N., Konteos, G., 2023. The impact of corporate governance attributes on environmental and social performance: the case of
European region excluding companies from the Eurozone. Bus. Strateg. Environ. 32 (6), 3489–3512.
Gonçalves, M.J.A., da Silva, A.C.F., Ferreira, C.G., 2022. The future of accounting: how will digital transformation impact the sector? Informatics 9 (1), 19.
González-Carrasco, I., Jiménez-Márquez, J.L., López-Cuadrado, J.L., Ruiz-Mezcua, B., 2019. Automatic detection of relationships between banking operations using
machine learning. Inf. Sci. 485, 319–346.
Goto, M., 2021. Collective professional role identity in the age of artificial intelligence. J. Professions Organ. 8 (1), 86–107.
Gray, G.L., Chiu, V., Liu, Q., Li, P., 2014. The expert systems life cycle in AIS research: what does it mean for future AIS research? Int. J. Account. Inf. Syst. 15 (4),
423–451.
Gregor, S., Benbasat, I., 1999. Explanations from intelligent systems: theoretical foundations and implications for practice. MIS Q. 497–530.
Grün, B., Hornik, K., 2011. topicmodels: an R package for fitting topic models. J. Stat. Softw. 40, 1–30.
Gunawardana, K.D., 2021. An analysis of financial distress prediction of selected listed companies in colombo stock exchange. Int. J. Sociotechnol. Knowl. Develop. 13
(2), 48–70.
Gupta, A., Sharma, U., Gupta, S.K., 2021. The role of ESG in sustainable development: an analysis through the lens of machine learning. In: 2021 IEEE International
Humanitarian Technology Conference (IHTC). IEEE, pp. 1–5.
Hajek, P., Henriques, R., 2017. Mining corporate annual reports for intelligent detection of financial statement fraud – a comparative study of machine learning
methods. Knowl.-Based Syst. 128, 139–152.
Hamal, S., Senvar, Ö., 2021. Comparing performances and effectiveness of machine learning classifiers in detecting financial accounting fraud for Turkish SMEs. Int. J.
Comput. Intell. Syst. 14 (1), 769–782.
Han, C., 2022.. Research on the application of artificial intelligence technology in accounting reform. In: 2022 IEEE Asia-Pacific Conference on Image Processing,
Electronics and Computers (IPEC). IEEE, pp. 238–244.
Hanauer, M.X., Kalsbach, T., 2023. Machine learning and the cross-section of emerging market stock returns. Emerg. Mark. Rev. 55, 101022.
Handoko, B.L., Liusman, S., 2021. Analysis of external auditor intentions in adopting artificial intelligence as fraud detection with the unified theory of acceptance
and use of technology (UTAUT) approach. In: The 2021 12th International Conference on E-business, Management and Economics (pp. 96-103).
Handoko, B.L., Mulyawan, A.N., Samuel, J., Rianty, K.K., Fgunaawa, S., 2019. Facing industry revolution 4.0 for millennial accountants. Int. J. Innov. Technol. Explor.
Eng. 9 (1), 1037–1041.
Hasan, A.R., 2021. Artificial Intelligence (AI) in accounting & auditing: a literature review. Open J. Business Manage. 10 (1), 440–465.
Hasan, M.M., Taylor, G., Richardson, G., 2022. Brand capital and stock price crash risk. Manag. Sci. 68 (10), 7221–7247.
Hastie, T., Tibshirani, R., Friedman, J.H., Friedman, J.H., 2009. The elements of statistical learning. In: Data Mining, Inference, and Prediction, vol. 2. Springer, New
York, pp. 1–758.
Henage, R., 2020. KPMG spark: bringing cutting-edge technology to SME clients. Acad. Acc. Financ. Stud. J. 24 (3), 1–7.
Hilb, M., 2020. Toward artificial governance? The role of artificial intelligence in shaping the future of corporate governance. J. Manag. Gov. 24, 851–870.
Hilmola, O.P., Gupta, M., 2015. Throughput accounting and performance of a manufacturing company under stochastic demand and scrap rates. Expert Syst. Appl. 42
(22), 8423–8431.
Hofmann, T., 1999. Probabilistic latent semantic indexing. In: Proceedings of the 22nd annual international ACM SIGIR Conference on Research and Development in
Information Retrieval (pp. 50–57).

15
B. Murphy et al. International Journal of Accounting Information Systems 55 (2024) 100709

Holmes, A.F., Douglass, A., 2022. Artificial intelligence: reshaping the accounting profession and the disruption to accounting education. J. Emerging Technol.
Account. 19 (1), 53–68.
Holowczak, R., Louton, D., Saraoglu, H., 2019. Testing market response to auditor change filings: a comparison of machine learning classifiers. J. Finance Data Sci. 5
(1), 48–59.
Hou, X., 2022. Design and application of intelligent financial accounting model based on knowledge graph. Mob. Inf. Syst.
Hsu, H.P., Hsu, H.M., 2008. Systematic modeling and implementation of a resource planning system for virtual enterprise by predicate/transition net. Expert Syst.
Appl. 35 (4), 1841–1857.
Huang, L., Yu, W., Ma, W., Zhong, W., Feng, Z., Wang, H., Chen, Q., Peng, W., Feng, X., Qin, B., Liu, T., 2023. A survey on hallucination in large language models:
principles, taxonomy, challenges, and open questions. arXiv preprint, arXiv:2311.05232.
Huang, F., No, W.G., Vasarhelyi, M.A., Yan, Z., 2022. Audit data analytics, machine learning, and full population testing. J. Finance Data Sci. 8, 138–144.
Hutchison, P.D., White, C.G., Daigle, R.J., 2004. Advances in accounting information systems and international journal of accounting information systems: first ten
volumes (1992–2003). Int. J. Account. Inf. Syst. 5 (3), 341–365.
Hyde, S.J., Bachura, E., Bundy, J., Gretz, R.T., Sanders, W.G., 2024. The tangled webs we weave: examining the effects of CEO deception on analyst recommendations.
Strateg. Manag. J. 45, 66–112.
Issa, H., Kogan, A., 2014. A predictive ordered logistic regression model as a tool for quality review of control risk assessments. J. Inf. Syst. 28 (2), 209–229.
Issa, H., Sun, T., Vasarhelyi, M.A., 2016. Research ideas for artificial intelligence in auditing: the formalization of audit and workforce supplementation. J. Emerging
Technol. Account. 13 (2), 1–20.
Jayasinghe, K., 2021. Constructing constructivism in management accounting education: reflections from a teaching cycle with innovative learning elements. Qual.
Res. Account. Manag. 18 (2), 282–309.
Jiang, Y., Li, X., Luo, H., Yin, S., Kaynak, O., 2022. Quo vadis artificial intelligence? Discover Artif. Intell. 2 (1), 4.
Jin, Y., 2024. Security monitoring method of accounting computerized software with artificial intelligence integration under digital marketing.
Jones, S., Alam, N., 2019. A machine learning analysis of citation impact among selected Pacific Basin journals. Acc. Finance 59 (4), 2509–2552.
Jørgensen, R.K., Igel, C., 2021. Machine learning for financial transaction classification across companies using character-level word embeddings of text fields. Intell.
Syst. Acc. Finance Manage. 28 (3), 159–172.
Juszczyk, M., Leśniak, A., 2019. Modelling construction site cost index based on neural network ensembles. Symmetry 11 (3), 411.
Kaabi, H., 2022. Comparative analysis of multicriteria inventory classification models for ABC analysis. Int. J. Inf. Technol. Decis. Mak. 21 (05), 1617–1646.
Kaplan, A., Haenlein, M., 2019. Siri, Siri, in my hand: Who’s the fairest in the land? On the interpretations, illustrations, and implications of artificial intelligence. Bus.
Horiz. 62 (1), 15–25.
Karaoglan, A.D., Karademir, O., 2017. Flow time and product cost estimation by using an artificial neural network (ANN): a case study for transformer orders. Eng.
Econ. 62 (3), 272–292.
Khalil, M.A., Zainuddin, Y., 2015. Intrinsic motivation as a mediator on accounting information system adoption. Pertanika J. Soc. Sci. Humanities 23, 33–46.
Khataie, A.H., Bulgak, A.A., Segovia, J.J., 2011. Activity-based costing and management applied in a hybrid decision support system for order management. Decis.
Support Syst. 52 (1), 142–156.
Kocaarslan, B., Soytas, U., 2023. The role of major markets in predicting the US municipal green bond market performance: new evidence from machine learning
models. Technol. Forecast. Soc. Chang. 196, 122820.
Kocsis, D., 2019. A conceptual foundation of design and implementation research in accounting information systems. Int. J. Account. Inf. Syst. 34, 100420.
Koh, S.L., Genovese, A., Acquaye, A.A., Barratt, P., Rana, N., Kuylenstierna, J., Gibbs, D., 2013. Decarbonising product supply chains: design and development of an
integrated evidence-based decision support system–the supply chain environmental analysis tool (SCEnAT). Int. J. Prod. Res. 51 (7), 2092–2109.
Korhonen, T., Selos, E., Laine, T., Suomala, P., 2021. Exploring the programmability of management accounting work for increasing automation: an interventionist
case study. Acc. Audit. Account. J. 34 (2), 253–280.
Kotseruba, I., Tsotsos, J.K., 2016. A review of 40 years of cognitive architecture research: core cognitive abilities and practical applications. arXiv preprint arXiv:
1610.08602.
Krippendorff, K., 1970. Estimating the reliability, systematic error and random error of interval data. Educ. Psychol. Meas. 30 (1), 61–70.
Krishna, S.H., Pattnaik, M., Patil, A., Rao, K.T.V., Muda, I., Saravanan, S., 2022. Artificial intelligence influence on accounting methods. In: 2022 5th International
Conference on Contemporary Computing and Informatics (IC3I). IEEE, pp. 2120–2125.
Kroon, N., do Céu Alves, M., Martins, I., 2021. The impacts of emerging technologies on accountants’ role and skills: connecting to open innovation—a systematic
literature review. J. Open Innov.: Technol. Market Complexity 7 (3), 163.
Kumar, K., Bhattacharya, S., 2006. Artificial neural network vs linear discriminant analysis in credit ratings forecast: a comparative study of prediction performances.
Rev. Acc. Financ. 5 (3), 216–227.
Kureljusic, M., Karger, E., 2023. Forecasting in financial accounting with artificial intelligence – a systematic literature review and future research agenda. J. Appl.
Acc. Res.
Lecue, F., Wu, J., 2017. Explaining and predicting abnormal expenses at large scale using knowledge graph based reasoning. J. Web Semant. 44, 89–103.
Lee, J.H.M., Leung, K.L., 2012. Consistency techniques for flow-based projection-safe global cost functions in weighted constraint satisfaction. J. Artif. Intell. Res. 43,
257–292.
Lehner, O., Leitner-Hanetseder, S., Eisl, C., 2019. The whatness of digital accounting: status quo and ways to move forward. ACRN J. Finance Risk Perspect. 8(2), pp.I-
X.
Lehner, O.M., Ittonen, K., Silvola, H., Ström, E., Wührleitner, A., 2022. Artificial intelligence based decision-making in accounting and auditing: ethical challenges and
normative thinking. Acc. Audit. Account. J. 35 (9), 109–135.
Leitner-Hanetseder, S., Lehner, O.M., Eisl, C., Forstenlechner, C., 2021. A profession in transition: actors, tasks and roles in AI-based accounting. J. Appl. Acc. Res. 22
(3), 539–556.
Lesner, C., Ran, A., Rukonic, M., Wang, W., 2020. Large scale personalized categorization of financial transactions. AI Mag. 41 (3), 63–77.
Li, J., 2019. Government accounting optimization based on computational linguistics. Cogn. Syst. Res. 57, 85–91.
Li, C. and Liu, X., 2022. Artificial neural networks in intelligent management systems for corporate financial accounting. In: 2022 International Conference on
Artificial Intelligence of Things and Crowdsensing (AIoTCs) (pp. 106–110). IEEE.
Li, J., 2022a. Construction and model realization of financial intelligence system based on multisource information feature mining. Comput. Intell. Neurosci.
Li, J., 2022b. An early control algorithm of corporate financial risk using artificial neural networks. Mobile Inf. Syst.
Liang, T.P., Chandler, J.S., Han, I., Roan, J., 1992. An empirical investigation of some data effects on the classification accuracy of probit, ID3, and neural networks.
Contemp. Account. Res. 9 (1), 306–328.
Lin, T.H., 2009. A cross model study of corporate financial distress prediction in Taiwan: multiple discriminant analysis, logit, probit and neural networks models.
Neurocomputing 72 (16–18), 3507–3516.
Lin, H.Y., 2021, November. Applying discriminant analysis and LVQ neural network to an empirical study of corporate social responsibility on Taiwanese electronics
companies. In: 2021 International Conference on Technologies and Applications of Artificial Intelligence (TAAI) (pp. 188–193). IEEE.
Liou, F.M., 2008. Fraudulent financial reporting detection and business failure prediction models: a comparison. Manag. Audit. J. 23 (7), 650–662.
Liu, C., Ryan, D., Lin, G., Xu, C., 2023. No rose without a thorn: Corporate teamwork culture and financial statement misconduct. J. Behav. Exp. Financ. 37, 100786.
Lokanan, M.E., 2019. A fraud investigation plan for a false accounting and theft case. J. Financial Crime 26 (4), 1216–1228.
Lynn, T., Endo, P.T., Rosati, P., Silva, I., Santos, G.L., Ging, D., 2019. A comparison of machine learning approaches for detecting misogynistic speech in urban
dictionary. In: 2019 International Conference on Cyber Situational Awareness, Data Analytics And Assessment (Cyber SA) (pp. 1-8). IEEE.
Mahlendorf, M.D., Martin, M.A., Smith, D., 2023. Innovative data–use-cases in management accounting research and practice. Eur. Acc. Rev. 1–30.
Makridakis, S., 2017. The forthcoming Artificial Intelligence (AI) revolution: its impact on society and firms. Futures 90, 47–60.

16
B. Murphy et al. International Journal of Accounting Information Systems 55 (2024) 100709

Malhotra, A., 2023. PSO based capital accounting information disclosure model using machine learning approach for smart index system. In: 2023 9th International
Conference on Advanced Computing and Communication Systems (ICACCS) (Vol. 1, pp. 2408-2412). IEEE.
Mancini, D., Lombardi, R., Tavana, M., 2021. Four research pathways for understanding the role of smart technologies in accounting. Meditari Acc. Res. 29 (5),
1041–1062.
Manita, R., Elommal, N., Baudier, P., Hikkerova, L., 2020. The digital transformation of external audit and its impact on corporate governance. Technol. Forecast. Soc.
Chang. 150, 119751.
Mardini, G.H., Alkurdi, A., 2021. Artificial intelligence literature in accounting: a panel systematic approach. In: The Fourth Industrial Revolution: Implementation of
Artificial Intelligence for Growing Business Success. Springer International Publishing, Cham, pp. 311–323.
Marques, S., Gonçalves, R., da Costa, R.L., Pereira, L.F., Dias, A.L., 2023. The impact of intelligent systems on management accounting. Int. J. Intell. Inf. Technol. 19
(1), 1–32.
Marrone, M., Hazelton, J., 2019. The disruptive and transformative potential of new technologies for accounting, accountants and accountability: a review of current
literature and call for further research. Meditari Acc. Res. 27 (5), 677–694.
Marshall, T.E., Lambert, S.L., 2018. Cloud-based intelligent accounting applications: accounting task automation using IBM watson cognitive computing. J. Emerg.
Technol. Account. 15 (1), 199–215.
Meservy, R.D., Denna, E.L., Hansen, J.V., 1992. Application of artificial intelligence to accounting, tax, and audit services: research at Brigham Young University.
Expert Syst. Appl. 4 (2), 213–218.
Minhas, S., Hussain, A., 2016. From spin to swindle: identifying falsification in financial text. Cogn. Comput. 8, 729–745.
Moepya, S.O., Akhoury, S.S., Nelwamondo, F.V., Twala, B., 2016. The role of imputation in detecting fraudulent financial reporting. Int. J. Innov. Comput. Inf. Control
12 (1), 333–356.
Moll, J., Yigitbasioglu, O., 2019. The role of internet-related technologies in shaping the work of accountants: New directions for accounting research. Br. Account.
Rev. 51 (6), 100833.
Moore, W.B., Felo, A., 2022. The evolution of accounting technology education: analytics to STEM. J. Educ. Bus. 97 (2), 105–111.
Moro, S., Cortez, P., Rita, P., 2015. Business intelligence in banking: a literature analysis from 2002 to 2013 using text mining and latent Dirichlet allocation. Expert
Syst. Appl. 42 (3), 1314–1324.
Muehlmann, B.W., Chiu, V., Liu, Q., 2015. Emerging technologies research in accounting: JETA’s first decade. J. Emerg. Technol. Account. 12 (1), 17–50.
Munoz, J., Jalili, M., Tafakori, L., 2022. Hierarchical classification for account code suggestion. Knowl.-Based Syst. 251, 109302.
Murphy, B., Feeney, O., 2023. AI, data analytics and the professions. In: The Future of Work: Challenges and Prospects for Organisations, Jobs and Workers. Springer
International Publishing, Cham, pp. 35–51.
Mushtaq, R., Gull, A.A., Shahab, Y., Derouiche, I., 2022. Do financial performance indicators predict 10-K text sentiments? An application of artificial intelligence.
Res. Int. Bus. Financ. 61, 101679.
Nado, R., Chams, M., Delisio, J., Hamscher, W., 1996. COMET: An application of model-based reasoning to accounting systems. AI Mag. 17 (4), 55.
Nielsen, S., 2022. Management accounting and the concepts of exploratory data analysis and unsupervised machine learning: a literature study and future directions.
J. Acc. Organ. Chang. 18 (5), 811–853.
Noble, W.S., 2006. What is a support vector machine? Nat. Biotechnol. 24 (12), 1565–1567.
Nóbrega, V., Costa, R.L.D., Gonçalves, R., Dias, Á., Pereira, L., Dorner, K., 2023. The impact of artificial intelligence in accounting: application in SMEs. Int. J.
Electron. Finance 12 (2), 192–214.
O’Leary, D.E., 2010. On the number of ISI citations to JETA, JIS, IJAIS, and ISAFM. J. Emerging Technol. Account. 7 (1), 89–100.
Odeyinka, H.A., Lowe, J., Kaka, A.P., 2013. Artificial neural network cost flow risk assessment model. Constr. Manag. Econ. 31 (5), 423–439.
Oesterreich, T.D., Teuteberg, F., 2019. The role of business analytics in the controllers and management accountants’ competence profiles: an exploratory study on
individual-level data. J. Account. Organ. Chang. 15 (2), 330–356.
O’Leary, D.E., 1991. Artificial intelligence and expert systems in accounting databases: survey and extensions. Expert Syst. Appl. 3 (1), 143–152.
Omar, N., Johari, Z.A., Smith, M., 2017. Predicting fraudulent financial reporting using artificial neural network. J. Financial Crime 24 (2), 362–387.
Östermark, R., 2015. A parallel algorithm for optimizing the capital structure contingent on maximum value at risk. Kybernetes 44 (3), 384–405.
Page, D., Seetharam, Y., Auret, C., 2023. Machine learning and manager selection: evidence from South Africa. Int. J. Emerg. Markets.
Papík, M., Papíková, L., 2022. Detecting accounting fraud in companies reporting under US GAAP through data mining. Int. J. Account. Inf. Syst. 45, 100559.
Papík, M., Papíková, L., 2023. Impacts of crisis on SME bankruptcy prediction models’ performance. Expert Syst. Appl. 214, 119072.
Perols, J., 2011. Financial statement fraud detection: an analysis of statistical and machine learning algorithms. Auditing: A J. Practice Theory 30(2), pp.19-5, pp.384-
405.
Petkov, R., 2020. Artificial intelligence (AI) and the accounting function – a revisit and a new perspective for developing framework. J. Emerging Technol. Account.
17 (1), 99–105.
Puthukulam, G., Ravikumar, A., Sharma, R.V.K., Meesaala, K.M., 2021. Auditors’ perception on the impact of artificial intelligence on professional skepticism and
judgment in Oman. Universal J. Acc. Finance 9 (5), 1184–1190.
Qasaimeh, G., Al-Gasaymeh, A., Kaddumi, T., Kilani, Q., 2022. Expert systems and neural networks and their impact on the relevance of financial information in the
Jordanian commercial banks. In: 2022 International Conference on Business Analytics for Technology and Security (ICBATS). IEEE, pp. 1–7.
Ragothaman, S., Davies, T., Dykstra, D., 2000. Legal aspects of electronic commerce and their implications for the accounting profession. Hum. Syst. Manag. 19 (4),
245–254.
Ranta, M., Ylinen, M., Järvenpää, M., 2023. Machine learning in management accounting research: Literature review and pathways for the future. Eur. Acc. Rev. 32
(3), 607–636.
Reepu, 2020. Role of artificial intelligence in finance and accounting. Int. J. Adv. Sci. Technol. 29 (4), 2275–2281.
Rest, J.R., 1986. Moral Development: Advances in Research and Theory. Praeger, New York.
Rikhardsson, P., Yigitbasioglu, O., 2018. Business intelligence & analytics in management accounting research: Status and future focus. Int. J. Account. Inf. Syst. 29,
37–58.
Rodrigues, L., Pereira, J., da Silva, A.F., Ribeiro, H., 2023. The impact of artificial intelligence on audit profession. J. Inf. Syst. Eng. Manage. 8 (1).
Russell, S., Norvig, P., 2010. Artificial Intelligence: A Modern Approach, third ed. University of California, Berkeley.
Salameh, R., Lutfi, K., 2021. The role of artificial intelligence on limiting Jordanian commercial banks cybercrimes. Accounting 7 (5), 1147–1156.
Salton, G., Buckley, C., 1988. Term-weighting approaches in automatic text retrieval. Inf. Process. Manag. 24 (5), 513–523.
Samanthi, D., Gooneratne, T., 2023. Bean counter to value-adding business partner: the changing role of the accountant and situated rationality in a multinational
firm. J. Account. Organ. Chang. 19 (3), 513–535.
Sánchez-Medina, A.J., Blázquez-Santana, F., Alonso, J.B., 2019. Do auditors reflect the true image of the company contrary to the clients’ interests? An artificial
intelligence approach. J. Bus. Ethics 155, 529–545.
Shao, J., Lai, K.K., Zheng, P., Zhang, G., Qin, Y., Lu, W., 2022. Enterprise accounting information identification and strategic management under data mining
technology. Mobile Inf. Syst.
Shen, W., Wu, Q., Chen, S., 2021. Enterprise intelligent reimbursement system based on OCR technology and SVM algorithm. In: 2021 4th International Conference on
Information Systems and Computer Aided Education (pp. 2639–2646).
Siew, E.G., Rosli, K., Yeow, P.H., 2020. Organizational and environmental influences in the adoption of computer-assisted audit tools and techniques (CAATTs) by
audit firms in Malaysia. Int. J. Account. Inf. Syst. 36, 100445.
Škapa, S., Bočková, N., Doubravský, K., Dohnal, M., 2023. Fuzzy confrontations of models of ESG investing versus non-ESG investing based on artificial intelligence
algorithms. J. Sustain. Finance Invest. 13 (1), 763–775.

17
B. Murphy et al. International Journal of Accounting Information Systems 55 (2024) 100709

Skrbiš, Z., Laughland-Booÿ, J., 2019. Technology, change, and uncertainty: maintaining career confidence in the early 21st century. N. Technol. Work. Employ. 34
(3), 191–207.
Sohangir, S., Wang, D., 2017. Improved sqrt-cosine similarity measurement. J. Big Data 4 (1), 1–13.
Solikin, I., Darmawan, D., 2023. Impact of artificial intelligence in improving the effectiveness of accounting information systems. J. Wireless Mobile Networks
Ubiquitous Comput. Dependable Appl. 14 (2), 82–93.
Spear, N.A., Leis, M., 1997. Artificial neural networks and the accounting method choice in the oil and gas industry. Account. Manag. Inf. Technol. 7 (3), 169–181.
Sumaryati, A., Novitasari, E.P., Machmuddah, Z., 2020. Accounting information system, internal control system, human resource competency and quality of local
government financial statements in Indonesia. J. Asian Finance Econ. Business 7 (10), 795–802.
Sun, T., 2019. Applying deep learning to audit procedures: an illustrative framework. Account. Horiz. 33 (3), 89–109.
Sun, M., Urquía-Grande, E., Chamizo-González, J., Campo, C.D., 2022. Accounting earnings and economic growth, trends, and challenges: a bibliometric approach.
Comput. Intell. Neurosci. 2022.
Sun, T., Vasarhelyi, M.A., 2018. Embracing textual data analytics in auditing with deep learning. Int. J. Digit. Account. Res. 18, 49–67.
Sutton, S.G., Holt, M., Arnold, V., 2016. “The reports of my death are greatly exaggerated”—Artificial intelligence research in accounting. Int. J. Account. Inf. Syst. 22,
60–73.
Sutton, S.G., Arnold, V., Holt, M., 2018. How much automation is too much? Keeping the human relevant in knowledge work. J. Emerging Technol. Account. 15 (2),
15–25.
Ta, Q., Nguyen, T., 2020. Impact of factors on the quality of accounting information system in civil engineering construction enterprises. Manage. Sci. Lett. 10 (13),
2985–2992.
Taipaleenmäki, J., Ikäheimo, S., 2013. On the convergence of management accounting and financial accounting–the role of information technology in accounting
change. Int. J. Account. Inf. Syst. 14 (4), 321–348.
Tater, T., Gantayat, N., Dechu, S., Jagirdar, H., Rawat, H., Guptha, M., Gupta, S., Strak, L., Kiran, S., Narayanan, S., 2022. AI driven accounts payable transformation.
In: Proceedings of the AAAI Conference on Artificial Intelligence 36(11), pp. 12405–12413.
Tiberius, V., Hirth, S., 2019. Impacts of digitization on auditing: a Delphi study for Germany. J. Int. Account. Audit. Tax. 37, 100288.
Tiron-Tudor, A., Deliu, D., 2022. Reflections on the human-algorithm complex duality perspectives in the auditing process. Qual. Res. Account. Manag. 19 (3),
255–285.
Trustorff, J.H., Konrad, P.M., Leker, J., 2011. Credit risk prediction using support vector machines. Rev. Quant. Finan. Acc. 36, 565–581.
Ulrich, P.S., Kratt, M., 2021. Adopting digital technologies in management accounting-empirical evidence from german SMEs. In: Proceedings of the Pacific Asia
Conference on Information Systems. pp. 1–8.
Utomo, D., Suhartono, E., Machmuddah, Z., 2020. The effect of accounting information systems to facilitate supply chain management in retail companies: evidence
from Indonesia. Int. J Sup. Chain. Mgt 9 (3), 863–870.
Vărzaru, A.A., 2022a. Assessing artificial intelligence technology acceptance in managerial accounting. Electronics 11 (14), 2256.
Vărzaru, A.A., 2022b. Assessing the impact of AI solutions’ ethical issues on performance in managerial accounting. Electronics 11 (14), 2221.
Vasarhelyi, M., Greenstein, M., 2003. Underlying principles of the electronization of business: a research agenda. Int. J. Account. Inf. Syst. 4 (1), 1–25.
Vial, G., 2019. Understanding digital transformation: a review and a research agenda. J. Strateg. Inf. Syst. 28 (2), 118–144.
Vu, N.T., Nguyen, N.H., Tran, T., Le, B.T., Vo, D.H., 2023. A LASSO-based model for financial distress of the Vietnamese listed firms: does the covid-19 pandemic
matter? Cogent Econ. Finance 11 (1), 2210361.
Wang, P., 2008. March. what do you mean by “AI”? AGI 171, 362–373.
Wang, X., Luo, X., Hu, Y., 2021. Enterprise accounting and financial risk analysis system based on decision tree and SVM. In: 2021 4th International Conference on
Information Systems and Computer Aided Education (pp. 2015–2018).
Wang, Y., Hao, X., 2023. Forecasting the real prices of crude oil: What is the role of parameter instability? Energy Econ. 117, 106483.
Warner, G., Wijesinghe, S., Marques, U., Badar, O., Rosen, J., Hemberg, E., O’Reilly, U.M., 2015. Modeling tax evasion with genetic algorithms. Econ. Gov. 16,
165–178.
Wu, Y., 2021. Discussion on the application mode of financial informatization in small and medium-sized enterprises based on deep learning algorithm. In: 2021 4th
International Conference on Information Systems and Computer Aided Education (pp. 2728–2732).
Yang, Y., Yang, J.Q., Bao, R., Zhan, D.C., Zhu, H., Gao, X.R., Xiong, H., Yang, J., 2023. Corporate relative valuation using heterogeneous multi-modal graph neural
network. IEEE Trans. Knowl. Data Eng. 35 (1), 211–224.
Yang, S., Zhu, S., 2023. Thinking and exploration of the teaching mode of empirical accounting course based on the Internet of Things and deep learning. Int. J. Netw.
Manag. e2242.
Yang, W., 2021. Cultivation of accounting professionals in private engineering colleges in the era of artificial intelligence–take GDLGXY as an example. In: 2021 4th
International Conference on Information Systems and Computer Aided Education (pp. 890–895).
Yang, N., 2022. Financial big data management and control and artificial intelligence analysis method based on data mining technology. Wireless Commun. Mobile
Comput.
Zand, A., Orwell, J., Pfluegel, E., 2020. A secure framework for anti-money-laundering using machine learning and secret sharing. In: 2020 International Conference
on Cyber Security and Protection of Digital Services (cyber Security). IEEE, pp. 1–7.
Zemankova, A., 2019. Artificial intelligence and blockchain in audit and accounting: literature review. WSEAS Trans. Business Econ. 16 (1), 568–581.
Zeng, H., 2022. Influences of mobile edge computing-based service preloading on the early-warning of financial risks. J. Supercomput. 78 (9), 11621–11639.
Zeng, Y., 2022b. Neural network technology-based optimization framework of financial and management accounting model. Comput. Intell. Neurosci.
Zengul, F., Bulut, A., Oner, N., Ahmed, A., Yadav, M., Gray, H.G., Ozaydin, B., 2023. A practical and empirical comparison of three topic modeling methods using a
COVID-19 corpus: LSA, LDA, and Top2Vec. In: Proceedings of the 56th Hawaii International Conference on System Sciences, Hawaii, USA.
Zhang, W., 2023. The application of cloud accounting in enterprise financial decision making in the era of big data. Appl. Math. Nonlinear Sci.
Zhang, W., Zhu, M., 2022. Environmental accounting system model based on artificial intelligence blockchain and embedded sensors. Comput. Intell. Neurosci.
Zhang, C.A., Cho, S., Vasarhelyi, M., 2022. Explainable Artificial Intelligence (XAI) in auditing. Int. J. Account. Inf. Syst. 46, 100572.
Zhang, Y., Li, Y., Cui, L., Cai, D., Liu, L., Fu, T., Huang, X., Zhao, E., Zhang, Y., Chen, Y. and Wang, L., 2023a. Siren’s song in the AI ocean: a survey on hallucination in
large language models. arXiv preprint, arXiv:2309.01219.
Zhang, C., Liang, D., 2023. Data integrity verification algorithm of accounting informatization cloud based on genetic optimization neural network. In: 2023 Asia-
Europe Conference on Electronics, Data Processing and Informatics (ACEDPI). IEEE, pp. 68–72.
Zhang, X., Su, J., Zhou, L., 2021. Fraud identification of financial reports based on neural network algorithm. In: 2021 40th Chinese Control Conference (CCC). IEEE,
pp. 8045–8049.
Zhang, C., Zhu, W., Dai, J., Wu, Y., Chen, X., 2023b. Ethical impact of artificial intelligence in managerial accounting. Int. J. Account. Inf. Syst. 49, 100619.
Zhao, N., Yen, D.C., Chang, I.C., 2004. Auditing in the e-commerce era. Inf. Manag. Comput. Secur. 5, 389–400.
Zhao, J., Zhang, L., Zhao, Y., 2022. Informatization of accounting systems in small-and medium-sized enterprises based on artificial intelligence-enabled cloud
computing. Comput. Intell. Neurosci.
Zhou, H., 2021. Error correction method of enterprise product cost accounting based on machine learning algorithm. Int. J. Prod. Dev. 25 (2), 101–113.
Zhou, K., 2023. Financial model construction of a cross-border e-commerce platform based on machine learning. Neural Comput. Appl. 1–11.
Zhu, S., 2021. Audit and technology. In: 2021 3rd International Conference on Machine Learning, Big Data and Business Intelligence (MLBDBI). IEEE, pp. 745–749.
Zuin, G., Parreiras, L., Melo, L., Barros, G., Lomeu, H., Melo, B., Marini, W., Lott, D., De Souza, M., 2023. An ensemble approach for inconsistency detection in medical
bills: a case study. In: 2023 IEEE 36th International Symposium on Computer-Based Medical Systems (CBMS). IEEE, pp. 573–578.
Zupan, M., Budimir, V., Letinic, S., 2020. Journal entry anomaly detection model. Intell. Syst. Acc. Finance Manage. 27 (4), 19.

18