AI Assignment New
AI Assignment New
copyright protection?
Abstract
The focus of this dissertation is on the effect that AI has on the digital art market
and the resulting ramifications pertaining to copyright protection. In recent
years, AI tools and algorithms have set in motion revolutionary changes affecting
numerous creative sectors; traditional notions of authorship and originality are
being challenged and re-evaluated, hence demanding a comprehensive review of
the current legal framework. By using a combination of case law analysis and
assessment of secondary sources, this study explores how AI is transforming
creation, distribution and copyright protection.
It deals with such critical issues as the determination of authorship in AI-
generated works, the developing notion of originality within the context of
machine learning, and whether present copyright laws are adequate to deal with
these technological developments. The dissertation also examines AI's role in the
creative process, highlighting how it can serve as a powerful tool for artists to
enhance their work and gain inspiration.
Table of Contents
Abstract.............................................................................................................................1
Chapter 1...........................................................................................................................1
1.1 Introduction.................................................................................................................1
1.11 Background............................................................................................................1
1.12 Keywords and Context...........................................................................................1
1.13 Research Significance.............................................................................................2
1.14 Objectives and Aims...............................................................................................2
1.15 Structure of The Dissertation.................................................................................3
1.2 Literature Review.........................................................................................................3
1.21 Structure of Literature review................................................................................3
1.22 AI-Generated Works and the Challenges of Copyright.....................................4
1.23 Arguments for Extending Copyright Protection to AI-Generated Works.........5
1.24 Arguments Against Extending Copyright Protection to AI-Generated Works. 6
1.25 Comparative Perspectives on AI and Copyright................................................6
1.26 The Role of Human Involvement in AI-Generated Works................................7
1.27 Gaps in the Literature........................................................................................8
1.3 Methodology...........................................................................................................8
1.31 Outline...............................................................................................................8
1.32 Research Philosophy..........................................................................................8
1.33 Research Approach............................................................................................9
1.34 Research Design...............................................................................................10
1.35 Research Limitations........................................................................................11
Chapter 2.........................................................................................................................12
2.1 AI in Creative Processes.........................................................................................12
2.11 How AI Models Operate...................................................................................12
2.12 Revolutionizing the Creative Process: How AI Benefits Artists.........................14
Chapter 3.........................................................................................................................15
3.1 Copyright Law: Foundation and Legal Frameworks...............................................15
3.11 Copyright Law: Overview and The Nature of Authorship.................................15
3.2 The Eligibility of AI for Copyright Protection: Authorship Laws in Different
Regions....................................................................................................................18
3.3 Further Cases and Judicial Precedent.................................................................24
3.4 AI and The Issue of Originality............................................................................25
Chapter 4.........................................................................................................................31
4.1 Policy recommendations........................................................................................31
2.41 Designating AI as Inventors..............................................................................31
2.42 Retaining the need for human authorship.......................................................32
2.43 Implementation of a new sui generis right for AI generated output................33
Chapter 5.........................................................................................................................35
5.1 Summary of Findings..............................................................................................35
5.2 Interpretation of Results........................................................................................35
5.3 Addressing the Research Questions.......................................................................35
5.4 Recommendations for Future Research.................................................................36
5.5 Practical Recommendations...................................................................................36
5.6 Concluding Remarks...............................................................................................36
Bibliography Table:..........................................................................................................37
1
Chapter 1
1.1 Introduction
1.11 Background
The advent and proliferation of artificial intelligence (AI) has resulted in a new
era of innovation, drastically transforming various aspects of human life,
including creative sectors such as the digital art market. This technological
evolution is constantly redefining and altering the boundaries of authorship,
creativity and ownership, presenting both unprecedented opportunities as well
as significant challenges. As AI technologies grow increasingly sophisticated, their
integration into the digital art landscape raises critical questions about the
nature of art itself and the tools that govern its creation and distribution.
Understanding the impact of AI on the digital art market necessitates a thorough
exploration of the intersection between technology, creativity, and legal
frameworks. This dissertation aims to examine the transformative impact of AI
on the digital art market and to analyze the implications for copyright protection.
1.12 Keywords and Context
To address the research question in a comprehensive manner, key terms and
concepts surrounding the issue at hand must be established.
efficient manner.1 These subsets of AI will be discussed in further length within
the literature review.
The following list is a summarization of the key terms and concepts prevalent
within the scope of the research question.
Keywords:
1
Smee S, ‘Perspective | AI Is No Threat to Traditional Artists. But It Is Thrilling.’ Washington Post (28
February 2023) <https://www.washingtonpost.com/arts-entertainment/2023/02/15/ai-in-art/> accessed 31
July 2024
2
The intersection of AI and digital art is an emerging field which has significant
implications for artists, collectors, legal professionals, and policymakers. AI tools
are not only enhancing human creativity but also generating entirely
autonomous artworks, having a widespread transformative effect on the digital
art l1andscape.2 The process of AI art generation is carried out by numerous
algorithms which utilize and compile information from existing art styles and
imagery across the internet., the advent of AI tools has raised important
concerns about authorship, originality, and the application of copyright law
within the digital art sphere.3 This ongoing debate surrounding the legal status of
AI-generated art has significant ramifications for artists, collectors, and the
broader art community, shaping the development of fair AI models 4. With the
line between human and machine-generated art becoming increasingly blurred,
it is crucial to engage in a thoughtful and nuanced discussion about the changing
nature of art and the role of AI in its future. By examining the legal challenges
and opportunities presented by AI-generated art, we can better understand the
profound impact these technological advancements will have on the
development of copyright law, as well as the art world and society as a whole.
2
Then C and others, ‘The Impact of Artificial Intelligence on Art - A Systematic Literature Review’
(2023) 2023 IEEE 9th Information Technology International Seminar (ITIS) (IEEE)
https://ieeexplore.ieee.org/document/10420208/ accessed 31 July 2024.
3
Ibid.
4
Yurii Burlyo, ‘AI-generated works and copyright protection’ (2022) 3 Entrepreneurship, Economy
and Law 7.
3
5
Joshua Yuvaraj, ‘Rethinking Originality in Copyright: The Impact of AI-Generated Works’ (2020)
41 European Intellectual Property Review 505.
5
The use of algorithms and datasets in AI-generated works raises questions about
the ownership of these underlying elements. For example, Andres Guadamuz
explores the implications of algorithmic creativity and the potential for copyright
infringement when AI systems are trained on copyrighted works without
permission. This issue is particularly contentious as it challenges the boundaries
of fair use and the rights of original content creators, who may find their work
used without attribution or compensation.6
Lionel Bently and Brad Sherman highlight the economic rationale behind
copyright protection, arguing that by granting exclusive rights to AI-generated
works, creators and companies are more likely to invest in AI technologies that
contribute to the cultural and economic landscape.7 The economic argument is
also supported by Michael P. Leahy, who contends that without copyright
protection, there would be less financial motivation for businesses to develop AI
capable of creating valuable content.8
Another argument in favor of extending copyright protection is the need for the
law to evolve in response to technological advances. Annemarie Bridy asserts that
copyright law has historically adapted to new forms of media and creativity, and
AI should be no exception. Bridy suggests that by recognizing AI-generated
works under copyright law, the legal system would reflect contemporary realities
and ensure that the law remains relevant in a rapidly changing technological
environment.9
6
Andres Guadamuz, ‘Artificial Intelligence and Copyright’ (2017) 18 WIPO Magazine 3.
7
Lionel Bently and Brad Sherman, Intellectual Property Law (5th edn, OUP 2018).
8
Ibid.
9
Annemarie Bridy, ‘Coding Creativity: Copyright and the Artificially Intelligent Author’ (2012) 5
Stanford Technology Law Review 1.
6
William Fisher argues that copyright law is fundamentally rooted in the protection
of human intellectual effort and extending it to AI-generated works would dilute
the essence of what copyright is meant to protect. Fisher’s perspective is echoed
by Jane C. Ginsburg, who asserts that copyright is intrinsically linked to human
authorship and moral rights which AI cannot fulfill.10
In the United States, the Copyright Office has maintained that only works created
by human authors are eligible for copyright protection. Sarah Burstein’s analysis
of U.S. copyright law underscores the strict adherence to the requirement of
human authorship, which has led to the rejection of copyright claims for AI-
generated works. However, recent discussions in legal circles suggest that there
may be a gradual shift towards recognizing AI’s role in the creative process,
though this is still a matter of significant debate.12
10
J C Ginsburg, ‘The Concept of Authorship in Comparative Copyright Law’ (2003) 52 DePaul Law
Review 1063, 1066
11
Timothy Endicott, ‘AI and Copyright: Overprotection Risks and Ethical Implications’ (2021) 34
International Journal of Law and Information Technology 77.
12
Sarah Burstein, ‘Are AI-Generated Works Copyrightable? An Analysis of U.S. Copyright Law’
(2019) 27 Journal of the Copyright Society of the USA 1.
7
The European Union has shown a more open stance towards the possibility of
extending copyright protection to AI-generated works. Niva Elkin-Koren
discusses the ongoing debates within the EU about how to best integrate AI into
existing intellectual property frameworks. She notes that while the EU is cautious
about fully recognizing AI as an author, there is a growing recognition of the need
to adapt copyright laws to accommodate the contributions of AI in creative
processes.13
In the UK, the Copyright, Designs and Patents Act 1988 includes provisions that
could potentially apply to AI-generated works, specifically Section 9(3), which
allows for copyright to be granted to the person who made the necessary
arrangements for the creation of a work if no human author can be identified.
Tanya Aplin examines this provision and its potential implications for AI-
generated works, arguing that it may provide a basis for recognizing copyright in
such works while still adhering to traditional copyright principles.14
Shlomit Yanisky-Ravid and Luis Javier Vasquez argue that if significant human
input is involved in the AI creative process—such as programming the AI,
curating the data sets, or refining the outputs—then copyright protection may be
justified. They propose a model where human creators who contribute to the
creative process can be recognized as authors, thereby maintaining the human-
centric nature of copyright while acknowledging the role of AI.15
13
Niva Elkin-Koren, ‘Copyright in a Digital Ecosystem: A User-Based Perspective’ (2017) 22
Journal of Intellectual Property Law 115.
14
Tanya Aplin, ‘Copyright in AI-Generated Works: A UK Perspective’ (2019) 42 Intellectual
Property Quarterly 25.
15
Shlomit Yanisky-Ravid and Luis Javier Vasquez, ‘In the Age of AI: New Models of Copyright
Authorship’ (2018) 24 Harvard Journal of Law & Technology 231.
8
Despite the growing body of literature on AI and copyright, several gaps remain.
Firstly, there is limited empirical research on the economic impact of extending
copyright protection to AI-generated works. While scholars like Bently and
Sherman have made theoretical arguments in favor of protection, there is a need
for empirical studies to assess the potential economic benefits and drawbacks.
Secondly, the ethical implications of granting copyright protection to AI works
have not been fully explored. The literature often focuses on legal and economic
arguments, leaving questions about the moral status of AI-generated works and
the potential consequences for human creators largely unexamined.
1.3 Methodology
1.31 Outline
The methodology section explains how the investigation into the intersection
between copyright and AI digital art has been approached in this dissertation.
The study chiefly relies on a comprehensive review of existing legal cases and
scholarly or journalistic literature to analyze key themes and issues. The research
design, instruments, philosophy and data collection methods will be detailed and
the reasoning behind choosing these procedures will also be elucidated. The
type of research conducted In this dissertation is secondary research.
16
Abbe E.L. Brown and Charlotte Waelde, ‘Authors, Machines, and Joint Works:
Reconceptualizing Copyright Law for the Digital Age’ (2017) 44 UC Davis Law Review 219.
17
Husam Helmi Alharahsheh and Abraham Pius, 'A Review of Key Paradigms: Positivism vs
Interpretivism' (2020) 2 Global Academic Journal of Humanities and Social Sciences 39-43.
18
Ibid.
9
legal principles and scholarly arguments in relation to copyright law and AI digital
art. This would be apt for making out complex legal interpretations and impacts
on society.19
The intersection of AI art and copyright protection is one that holds a myriad of
legal and ethical problems that are driven by technology and driven socially. In
this respect, complex issues like these would be best evaluated using
interpretivism20, since it allows for in-depth coverage of different stakeholders'
interpretations of these issues and their respective responses.
Copyright law and its application to AI art are fledgling areas of research that are
continuously evolving and progressing. Interpretivism aids in clarifying how legal
principles are being interpreted and adapted in reaction to new technological
realities. Due to the implications of the research question as well as the overall
importance and divisiveness of opinions on the dissertation topic, Different legal
scholars and practitioners may have varied interpretations of how copyright law
applies to AI art. Since the dissertation focuses on secondary research,
interpretivism supports a detailed analysis of existing scholarly articles, legal
texts, and case studies.21 This approach allows for the exploration of the diverse
perspectives scholars and legal practitioners hold on the topic.
the most plausible explanations or theories that can account for these
observations. In the context of AI art, copyright protection presents novel
challenges and questions that traditional legal theories do not fully address and
cover. For example, the legal status of AI-generated art, the implications of AI's
role in creative processes, and the adequacy of prevailing copyright frameworks
in managing these issues can be multifaceted and not very clear. 24
Present case law, and scholarly articles often offer fragmented or insufficient
outlooks on how copyright law should be applied to AI-generated art. An
abductive approach helps in making sense of these discrepancies and generating
explanations that account for them. This is evident in the literature review, when
examining global legal systems in correlation to legislation regarding authorship
and originality.25 By analyzing how numerous jurisdictions handle the issue of ai
art and copyright protection, emerging patterns were discerned relating to why
legal frameworks around the world have differing stances on what constitutes as
originality. Similarly, by identifying and quoting specific legal cases, the rulings
resulting from these cases were essential in recognizing what gaps and
limitations exist in the existing legal understanding of authorship, especially in
answering whether machines could classify as authors.26 Moreover, iterative
refinement is a critical component of an abductive approach. In this dissertation,
conclusions drawn of how case law influences copyright treatment of AI art has
been refined as new cases or additional data are considered. This style once
again reflects an abductive methodology.
relevant to the research topic, especially those affecting the artists and AI
developers themselves.
The research applies a purposive sampling strategy in the selection of legal cases
and literature material that best suits the focus of the dissertation on AI digital
art and copyright law. The criteria for selection centers on: relevance of AI-
created art to the issue of copyright, impact of AI on copyright issues, impact on
legal doctrines. This strategy entails deliberately choosing only the most relevant
sources that will lead to an effective analysis answering the research question
and achieving its objectives, contributing to a more focused analysis.
Chapter 2
2.1 AI in Creative Processes
2.11 How AI Models Operate
In order to explore the more widespread effects of AI on copyright legislation
and the ethical considerations present therein, the inner functioning of AI tools
must be elucidated. AI-generated content covers a wide array of creative works
produced with the aid of artificial intelligence systems. AI algorithms, employing
techniques such as machine learning, deep learning, and natural language
processing, have become increasingly adept at creating various forms of content,
including text, images, music, and even entire films.29 These algorithms analyze
extensive datasets, identify patterns, and generate content that emulates human
creativity. AI-generated content possesses unique characteristics that distinguish
it from traditional human-created works. While human creators draw from their
personal experiences, emotions, and cultural backgrounds, AI algorithms depend
on vast amounts of existing data to guide their decision-making processes. As
scholars emphasize “these systems draw upon extensive datasets, training their
models to produce innovative content by discerning and emulating established
patterns and artistic styles.”30
How exactly are these models trained and what procedure do they undergo?
Two of the most common types of AI models that are trained for creating art are
Generative Adversarial Networks (GANs), and Convolutional Neural Networks. 31
GANs consist of two neural networks: the generator and the discriminator. The
generator creates images, and the discriminator evaluates them against real
images from a training set. The generator gets feedback from the discriminator
about how real or fake its creations look and improves over time. This feedback
loop continues until the generator produces a result that is so realistic that the
29
Abdikhakimov, ‘Unraveling the Copyright Conundrum: Exploring AI-Generated Content and Its
Implications for Intellectual Property Rights’ (2023) 1(5) International Conference on Legal
Sciences 18–32 https://www.science-zone.org/index.php/conference/article/view/42 accessed
31 July 2024.
30
C Watiktinnakorn, J Seesai, and C Kerdvibulvech, ‘Blurring the Lines: How AI Is Redefining
Artistic Ownership and Copyright’ (2023) 3(1) Discover Artificial Intelligence 3, para 2.
31
J C Sacco and V Camilleri, ‘An Investigation into AI-Generated Art Through GANs and ML Neural
Network’ in K Arai (ed), Intelligent Computing (SAI 2024, Springer 2024) Lecture Notes in
Networks and Systems, vol 1016 https://doi.org/10.1007/978-3-031-62281-6_7 accessed 10 July
2024.
13
Style transfer is another prevalent technique in AI art where the style of one
image is applied to the content of another. This is achieved using convolutional
neural networks (CNNs). The AI, powered by a CNN, looks at both images and
breaks them down into layers of features.34 These features can be simple things
like edges and textures, or more complex patterns like swirls and colors. The AI
separates the content (shape and structure) from the style (the colors, brush
strokes, and patterns) of the art piece. It then blends these two sets of features:
it keeps the content of original piece intact while applying the artistic style of the
other artwork to it.35 This involves optimizing an image so that its content
matches the original photo or work, and its style matches the artistic image.
Workflow:
32
Ibid.
33
Jia Z and Li Y, ‘Digital Media Artistic Creation Based on Generation Confrontation Network and
Internet of Things’ [2023] Computer-Aided Design and Applications 150
https://cad-journal.net/files/vol_21/Vol21NoS13.html accessed 31 July 2024.
34
K B Bhangale, P Desai, S Banne, and U Rajput, ‘Neural Style Transfer: Reliving Art Through
Artificial Intelligence’ in 2022 3rd International Conference for Emerging Technology (INCET) (IEEE
2022) pp 1-6.
35
Jia and Li, ‘Digital Media Artistic Creation’ (n 1) 150.
36
E Cetinic and J She, ‘Understanding and Creating Art with AI: Review and Outlook’ (2022) 18(2)
ACM Transactions on Multimedia Computing, Communications, and Applications 1–22.
14
While the inner workings of AI reveal its powerful capacity to generate creative
outputs, they also expose significant challenges to existing legal frameworks,
particularly in the area of copyright. As AI continues to produce works that rival
human creativity, it challenges traditional notions of authorship, posing difficult
questions about how these creations should be protected—or whether they can be
protected at all—under current copyright law. The conventional view of copyright
law’s purpose, that it was established for the preservation of human creativity ,is
disrupted by AI outputs, as it is not entirely clear if the human input required in
AI generation is enough for copyright protection or sufficient to satisfy various
authorship standards. This emerging tension between AI innovation and legal
doctrine underscores the urgent need to reassess and potentially redefine
copyright principles, a topic that will be critically examined in the next chapter.
37
L Xian and E-W Huang, ‘Computer Generated Art: Exploring the Application of Artificial
Intelligence in Digital Art’ in System Innovation for a World in Transition (CRC Press 2023) pp 168–
172.
38
‘Embracing Creativity: How AI Can Enhance the Creative Process’ (sps.nyu.edu)
<https://www.sps.nyu.edu/homepage/emerging-technologies-collaborative/blog/2023/
embracing-creativity-how-ai-can-enhance-the-creative-process.html> accessed 31 July 2024
39
M A Ali Elfa and M E T Dawood, ‘Using Artificial Intelligence for Enhancing Human Creativity’
(2023) 2(2) Journal of Art, Design and Music 3.
40
Ibid.
15
Chapter 3
3.1 Copyright Law: Foundation and Legal Frameworks
3.11 Copyright Law: Overview and The Nature of Authorship
In order to examine the evolving relationship between artificial intelligence and
copyright laws, the context of the current legal system in relation to authorship
rights must be established. The copyright system encompasses laws that provide
for exclusive rights to copyrighted materials, and it was developed in the early
18th century when the printing press became more widely available 41. Specific
rights endowed in this regard include a work's duplicating or reproducing,
followed by publishing, presentation, and sharing before the public.42 To
determine who should be granted ownership of the copyright protection
attributed to AI artworks, we must first establish who the author of an AI
generated work is. This is because the author of the art is typically the owner of
the copyright protection accorded to the work, barring a few select cases. For
example, a company that commissions an AI-generated artwork might own the rights,
even if an individual programmer or artist is considered the author.43
In the field of creative works generated by AI, the first and most prevalent
conception of the term "author" refers to the originator of a piece.44 Some
scholars argue that AI should be more regarded as a tool or contrivance for a
human author rather than as an independent creator.45 Scholars have put
forward three candidates that could potentially serve as the author of an ai
generated work.46 Firstly, there is the human who coordinated the process of
creation, particularly in works that had assistance from AI.47 The attribution of
copyright to a human author vests in the theory that the presence of a
sufficiently close causal connection between what the human has contributed
and what the AI finally comes up could amount to an act of authorship.
Many commentators believe the second claimant to copyright authorship in this
regard is the AI program that generates the results themselves, especially in
situations where a work is exclusively created by an autonomous artificial
41
Burlyo, ‘AI-generated works’ (n 5).
42
WIPO Secretariat, ‘Revised Issues Paper on Intellectual Property Policy and Artificial
Intelligence’ (WIPO/IP/AI/2/GE/20/1 REV, 11 June 2022) para 23.
43
Ana Ramalho, ‘The Concept of Authorship and Ownership in AI-Generated Works: An
Intellectual Property Law Perspective’ (2020) 42 European Intellectual Property Review 773.
44
Ginsburg, ‘The Concept of Authorship’ (n 10) 1066.
45
Ibid.
46
WIPO Secretariat, ‘Revised Issues Paper on Intellectual Property Policy and Artificial
Intelligence’ (WIPO/IP/AI/2/GE/20/1 REV, 11 June 2022) para 23 (n 19).
47
E E Adaka and I A Olubiyi, ‘Lessons for Nigeria: Determining Authorship and Inventorship of
Artificial Intelligence Generated Works’ (2022) 2 Journal of Intellectual Property and Information
Technology Law 15.
16
Within the domain of literary and artistic content, the notion of authorship has
been based fundamentally on one aspect: human creativity. Hence, human
authors are given specific rights over their original creations. These rights
48
Ibid.
49
Daniel Gervais, ‘The Machine as Author’ (2020) 105(5) Iowa Law Review 2053.
50
Pamela Samuelson, ‘Allocating Ownership Rights in Computer-Generated Works’ (1986) 47(4)
University of Pittsburgh Law Review 1185.
51
A Guadamuz, ‘Do Androids Dream of Electric Copyright? Comparative Analysis of Originality in
Artificial Intelligence Generated Works’ (2017) 2 Intellectual Property Quarterly 169.
52
Ginsburg, ‘The Concept of Authorship’ (n 20) 1066.
53
Ibid.
17
empower them with the placing of their works on the market, gaining revenues
from such actions, making assignments of copyright, and safeguarding their
name from copyright infringement.54 The Berne Convention is considered a
landmark agreement in the field of copyright protection and represents cardinal
principles of a nature when one speaks about authors and their rights. This
international treaty gives general protection to all literary and artistic works and,
hence, offers its advantages to all authors and their successors.55 By doing this,
authors are empowered to protect and guard their creative works and claim
their rights. The Berne Convention further promotes clarity as it allows for the
clear indication of authors under their real name or pseudonym, therefore
completely ruling out any doubt concerning the authorship. In doing all this, the
Berne Convention plays a critical role in promoting protection, recognition, and
preservation of the works authored by people in literature and art. Basically, the
Berne Convention constitutes the core or host of other international instruments
that govern the understanding of authorship and copyright worldwide. 56
The development of copyright law has changed significantly over time. As stated,
It started with the invention of the printing press and has now reached a crucial
point with the rise of AI. This marks an important milestone in the evolution of
copyright law. Based on recent developments, global copyright systems have
been significantly disrupted by the outstanding capacity of AI technologies to
produce literary and artistic works in a consistently efficient manner. This
disruption is seen in ongoing debates about the problem of authorship in AI-
generated works.57 The key points of concern on which this discourse lies can be
summarized by statements attributed to Professor Sam Ricketson, who initially
asked the question of who the ‘rightful’ author behind artworks generated by AI
is. Did it stem from human creativity or the inventive capabilities of machines?
This insightful question resonates deeply with current debates surrounding
authorship in AI-generated content.58 Ricketson’s observation was that though
not specifically stating that an author had to be a human being, the Berne
Convention seemed to take this as an implied premise which its members shared
among themselves. This self-evident assumption poses some interesting
questions about how we see authors when regarding works produced by
artificial intelligence. The use of AI to produce literature and art has exploded
54
N Basri, ‘The Question of Authorship in Computer-Generated Work’ Journal of Law & Social
Change (13 January 2020) https://www.law.upenn.edu/live/news/9691-the-question-of-
authorship-in-computer-generated accessed 26 June 2024.
55
S Ricketson and J C Ginsburg, International Copyright and Neighboring Rights: The Berne
Convention and Beyond (2nd edn, Oxford University Press 2006) 87, 586.
56
Ibid.
57
A Bridy, ‘Coding Creativity: Copyright and the Artificially Intelligent Author’ (2012) Stanford
Technology Law Review 1, 24–27.
58
Ricketson and Ginsburg, International Copyright and Neighboring Rights (n 31) 586.
18
59
B E Boyden, ‘Emergent Works’ (2016) 39 Columbia Journal of Law & the Arts 377, 378–79; M E
Kaminski, ‘Authorship, Disrupted: AI Authors in Copyright and First Amendment Law’ (2017) 51
UC Davis Law Review 589, 592.
60
Ibid.
61
Copyright, Designs and Patents Act 1988, ss 77, 80
https://www.legislation.gov.uk/ukpga/1988/48/contents accessed 12 July 2024.
62
Ibid.
63
Ginsburg, ‘The Concept of Authorship’ (n 20) 1066.
64
Burrow-Giles Lithographic Co v Sarony 111 US 53 (1884).
19
created. The decision reflects the shift that is presently tearing through the
courts on copyright law as it relates to AI and other non-human authorship
issues.74
A number of prominent cases in this shifting legal landscape have further defined
the ambiguities where copyright law intersects with AI. One such lawsuit
involves the case of copyright infringement filed by Getty Images against Stability
AI for training its AI systems with copyrighted photos. According to reports,
Stability AI copied a minimum of 12 million copyrighted images that naturally
belong to Getty Images for its Generative AI models. In the UK, Getty Images
aimed at stopping the sales of Stability AI on grounds of having used the
copyrighted images to "train" their Stable Diffusion system.75 Also, A US federal
judge has dismissed an attempt to copyright artwork created by an AI image
generator, in a ruling that has underlined just how incredibly tricky the question
of authorship and copyright is when it comes to AI.76
In India, the law on copyright says that, in the case of computer-generated or
generated literary, dramatic, or artistic works, the person who causes them to be
created—arranges for their production—is considered as the author. It must be
noted at the same time that even when the contribution of AI in the process of
creation is very substantial, the human creator or programmer is regarded as
being the first and foremost cause of such works.77 There is no definition of
"author" under the heading of artistic and literary works within the Copyright Act
of 1957. However, the term is comprehensively and completely defined under
section 2(d) of the Act. The "author" in relation to a literary or dramatic work,
means the person who creates the work. For musical works, the "author" is the
composer. For artistic works excluding photographs, the "author" shall be the
artist. Moreover, according to this Act, the person taking a photograph is
considered an "author" with respect to photographs.78
What is most noteworthy, however, is the completely divergent approaches
Australia and Thailand have taken in the application of copyright law. Both
74
Franklin Graves, ‘U.S. Copyright Office Backtracks on Registration of Partially AI-Generated
Work’ (1 November 2022) IPWatchdog https://ipwatchdog.com/2022/11/01/us-copyright-office-
backtracks-registration-partially-ai-generated-work/id=152451/ accessed 16 July 2024.
75
Baba Tamim, ‘AI Image Generator Hit by $1.8 Trillion Lawsuit from Getty Images: The Stock
Image Agency Claims that Over 12 Million Copyrighted Images Along with Their Descriptions and
Metadata—Were Used to Train Stable Diffusion, an AI Image Creator’ (8 February 2023)
Interesting Engineering https://interestingengineering.com/culture/getty-images-lawsuit-against-
stability-ai accessed 17 July 2024.
76
‘As Fight over A.I. Artwork Unfolds, Judge Rejects Copyright Claim’ (21 August 2023) The New
York Times https://www.nytimes.com/2023/08/21/technology/ai-artwork-copyright.html
accessed 12 July 2024.
77
Indian Copyright Act 1957, s 2(d) (No 14 of 1957).
78
Ibid.
22
79
Nattapong Suwan-In, ‘Copyright Protection on AI-Generated Work: The Case Study of the US,
UK, and Thailand Copyright Laws’ (2021) Journal of Law Public Administration and Social Science
131, 142. See also Australian Copyright Act 1968 (Cth) (Copyright Act) and IceTV Pty Ltd v Nine
Network Australia Pty Ltd [2009] HCA 14.
80
Susanna HS Leong, ‘Legal Protection of Factual Compilations and Databases: Re-Thinking the
Copyright Protection Model in Singapore’ (2002) 5 Journal of World Intellectual Property 1047.
81
Asia Pacific Publishing Pte Ltd v Pioneers & Leaders (Publishers) Pte Ltd [2011] SGCA 37.
82
Evan Ng, ‘B2C2 Ltd v Quoine Pte Ltd [2019] SGHC (I) 03’ (2019) Singapore Commercial Law
Review 207.
83
Ibid.
23
These legal precedents make it clear that Singaporean law is not currently
providing overt protection for literary and artistic works created by AI. The
foregoing decisions bring forth the fact that there is no express provision of law
in this regard, and that there is a general reluctancy to afford copyright
protection to works that have been generated exclusively by AI tools.
The results from these countries show that there is a consistent belief in the
importance of human creativity as the main principle behind copyright law in
many parts of the world.
Currently, both case law and the practices of the US Copyright Office make it
clear that works created solely by AI without any human involvement are not
eligible for copyright protection. The main requirement for copyright protection
is that a human author must be involved in the creative process. This legal
framework reinforces the idea that copyright law is meant to protect the
originality and creative contributions of human creators. While AI technology has
increasingly evolved and can produce some impressive works, the fact alone that
no human involvement goes directly into the creative process could arguably
preclude AI generated digital art from gaining copyright protection. 84
3.3 Further Cases and Judicial Precedent
While there have been various cases assessing the aptness of AI systems for
copyright protection, to date, no case of copyright infringement of an AI work or
challenging the Human Authorship Requirement has been litigated in the United
States, and many other nations. Several factors may contribute to this. First of
all, copyright applicants may have gotten registrations by not disclosing that
their works were created by AI; similarly, the prosecution of claims for
infringement by not identifying the source of their works.85 More likely, the lack
of case law is indicative of the low value of AI-created works. For instance, while
an AI-created song might represent an impressive technical advance, if it is not
going to generate substantial streaming revenues, there is no reason to incur the
high legal fees involved in copyright litigation.86 The latest to make headlines—
and attract a copyright infringement threat—is Canadian artist Adam Basanta,
for his work entitled All We'd Ever Need Is One Another. This mixed-media
installation creates upwards of more than a thousand images per day using two
flatbed scanners set to face each other, running scanning cycles in perpetual
motion with randomized settings controlled by software.87 These images are
84
Gianmaria Ajani, ‘Human Authorship and Art Created by Artificial Intelligence – Where Do We
Stand?’ in Digital Ethics (Nomos Verlagsgesellschaft mbH & Co. KG 2022) 253–270.
85
Michael D Murray, ‘Generative AI Art: Copyright Infringement and Fair Use’ (2023) 26 SMU
Science and Technology Law Review 259.
86
Ibid.
87
Adam Basanta, ‘All We’d Ever Need is One Another’ (last accessed 18 July 2024)
http://allwedeverneed.com/.
24
Another important case is that of Gao Yang and Youku, decided by the Beijing
Court. It dealt with the copyrightability of photographs automatically taken by a
camera attached to a hot-air balloon. Thus, this was a case about whether AI-
generated photos could be protected under Chinese copyright law. A court ruled
that under China's legal outlines, the automatic photos captured by a plaintiff's
camera are entitled to copyright protection. Consequently, it was held that
defendant's use of the photos without authority infringed the plaintiff's
copyright in the photographs. What that decision drew to the attention of all was
88
Statement of Claim at para 30, quoted by Teresa Scassa, ‘Artist Sued in Canada for Copyright
Infringement for AI-Related Art Project’ (4 October 2018)
http://www.teresascassa.ca/index.php?option=com_k2&view=item&id=286 accessed 12 July
2024.
89
Shenzhen Tencent v Shanghai Yingxun Case No Y0305MC No 14010 (Nanshan District People’s
Court, Shenzhen, Guangdong, China 2019).
90
Ibid.
25
the fact that although taking the photo in question was automated by the hot-air
balloon itself, the plaintiff's placing of a sports camera was a sufficient basis for
creativity to merit copyright protection.91
The idea of originality in common law has traditionally been linked to the “sweat
of the brow” doctrine, but this standard for originality has changed in countries
such as US and is no longer necessary to establish authorship or qualify a work
for copyright. This approach suggests that its no longer customary for the extent
of mere labor carried out by the author to be the key deciding factor in whether
the work is deemed original, the conditions for attaining copyright protection are
now increasingly dependent on the creativity that was spent by the creator in
the conception of the work.96 Adding onto this view, the case of Hollinrake v
91
Beijing Intellectual Property Court Jing 73 Min Zhong No 797 Civil Judgment (2 April 2020).
92
Mark Fenwick and Paulius Jurcys, ‘Originality and the Future of Copyright in an Age of
Generative AI’ (2023) 51 Computer Law & Security Review 105892.
93
CCH Canadian Ltd v Law Society of Upper Canada [2004] SCC 13, para 16.
94
CCH Canadian Ltd v Law Society of Upper Canada [2004] SCC 13, para 16.
95
Hafiz Gaffar and Saleh Albarashdi, ‘Copyright Protection for AI-Generated Works: Exploring
Originality and Ownership in a Digital Landscape’ (2024) Asian Journal of International Law 1-24.
96
S Yanisky-Ravid and LA Velez-Hernandez, ‘Copyrightability of Artworks Produced by Creative
Robots, Driven by Artificial Intelligence Systems and the Concept of Originality: The Formality-
26
Truswell further expounds that the benchmark of originality is not reliant upon
the novelty, newness or innovation of a work but rather upon its link to the
author itself.97 This means that, in order for an artistic work to be labeled as
original, it must have been conceived and established by the author’s own
creativity, rather then originating as a facsimile of some other source. This
clarified that the originality of a work is derived from the creative skill utilized by
the author in creating it, and not its uniqueness in comparison to other works.
This is the standard legal classification of originality in jurisdictions such as Wales
and the UK.
102
Feist Publications Inc v Rural Telephone Service (n 42).
103
P Bernt Hugenholtz and João Pedro Quintais, 'Copyright and Artificial Creation: Does EU
Copyright Law Protect AI-Assisted Output?' (2021) 52(9) IIC-International Review of Intellectual
Property and Competition Law 1190, 1190-1216.
104
Ibid.
105
Football Dataco Ltd v Yahoo! UK Ltd [2012] EWCA Civ 1759, [2013] 1 All ER 15.
106
Brompton Bicycle Ltd v Chedech [2020] EWHC 104 (IPEC), [2020] FSR 25.
107
Cofemel – Sociedade de Vestuário SA v G-Star Raw CV (C-683/17) EU:C:2019:721.
28
Numerous academics and experts on the issue argue that the answer is a
definitive no. Professor Daniel Gervais states that originality is the core base
108
H B Abrams, 'Originality and Creativity in Copyright Law' (1992) 55 Law & Contemporary
Problems 3.
109
Gaffar and Albarashdi (n 71) 5.
110
Nahide Basri, ‘The Question of Authorship in Computer-Generated Work’ (13 January 2020)
Journal of Law & Social Change https://www.law.upenn.edu/live/news/9691-the-question-of-
authorship-in-computer-generated accessed 20 July 2024
29
upon which the idea of copyright protection is founded, and that AI generated
artistic content cannot be characterized as original because the kind of expansive
critical thinking that leads to novel individual expression is a feature solely of the
human mind.111 Moreover, he argues that for a work to be original and hence be
accredited to a specific author, the author must be able to assume responsibility
for his creations, which AIs, owing to their non-sentient nature, cannot do. 112
In the Third Session on Intellectual Property and Artificial Intelligence, the chief
dialogue focused on two opposing theories. Critics rejected the idea of granting
copyright protection to AI-created works and gave various justifications for such
a rejection.113 Some commentators highlighted that the technical input of human
factors together with AI algorithms would not justify the protection under
copyright of AI-created works. They explained that copyright law has always
insisted on human authorship.114 The meaning of AI was still ill-defined and
scantily developed. Further, they said that every work created by AI should be in
the public domain. They cautioned that the expansion in copyright to AI-assisted
creative works would entail policy problems at the stages of arrangement,
creation, and selection, and would deplete the public domain. It was their
argument that putting more AI-generated literary and artistic works under
copyright would restrict access of the public to such works.115
Chapter 4
4.1 Policy recommendations
2.41 Designating AI as Inventors
118
T Aplin and G Pasqualetto, ‘Artificial Intelligence and Copyright Protection’ in Rosa Maria
Ballardini, Petri Kuoppamäki and Olli Pitkänen (eds), Regulating Industrial Internet Through IPR,
Data Protection and Competition Law (Kluwer 2019).
119
Peter Georg Picht and Florent Thouvenin, ‘AI and IP: Theory to Policy and Back Again – Policy
and Research Recommendations at the Intersection of Artificial Intelligence and Intellectual
Property’ (2023) 54(6) IIC-International Review of Intellectual Property and Competition Law 916–
40.
31
The other step in this regard would be to integrate the AI capabilities into the
PHOSITA concept. It is a concept setting the level of creativity required to make
an invention patentable. The recognition of AI under such a concept will increase
the threshold of patentability, hence ensuring that the patents are granted to
works that are actually innovative and not a routine use of AI algorithms that
results in an insufficient level of inventiveness.
The idea of human authorship should be at the heart of copyright law, especially
in systems that follow droit d’auteur (author’s rights). This viewpoint emphasizes
that works created entirely by AI, without any human involvement, shouldn't be
eligible for copyright protection, even if they meet the standards for creativity
set by current copyright law.
As established, The main idea behind present copyright law is protecting and
promoting human creativity. Underlying this philosophy is the principle that
creative works represent an expression of the human mind, emotions, and
culture.122
By giving priority to human authorship, copyright law tends to provide for an
atmosphere whereby people are motivated to produce original works. 123
When a work is produced through the collaborative effort between an AI system
or tool and a human, the result may meet the standard for copyright protection.
120
Ibid.
121
Ibid.
122
E Lee, ‘Prompting Progress: Authorship in the Age of AI’ (2024) 76 Florida Law Review.
123
James Grimmelmann, ‘There’s No Such Thing as a Computer-Authored Work – And It’s a Good
Thing, Too’ (2016) 39 Columbia Journal of Law & the Arts 403, 403.
32
However, this is contingent upon the requirement that the human contribution
involved in the creation of the work must meet a certain threshold for creativity.
This contribution must exceed the rudimentary guidance or management of the
AI and play a notable role in the creative process that shapes the work. The work
must possess a significant degree of individual expression that is representative
of the creative output of the human mind. Hence, any piece of art that is
produced solely through an AI without the necessary human involvement and
effort required cannot qualify for copyright protection.124
2.43 Implementation of a new sui generis right for AI generated output
The established context underlies that there are many legal and ethical
uncertainties in relation to traditional copyright law when applied to ai
generated output. Hence, to achieve more clarity and transparency, a sui generis
right specifically tailored to AI produced artistic content would help mitigate
such uncertainties and lay a better groundwork for AI related legislation moving
forward. This viewpoint proclaims that, instead of enforcing the conventional
framework of copyright law onto AI generations, a new system of legal
protection altogether should be put into place to address the intellectual
property rights of AI works.
Moral rights should not be afforded to holders of sui generis rights for content
resulting from AI-generation.128 The U.S. Copyright Act supports moral rights,
such as attribution and integrity, under which creators of visual art may claim
authorship and preserve the personal and reputational value of their works.
These may include the right of the creator to be identified as the author of the
work, the right of the creator to prevent any distortion or other modifications of
their work, and the right of the creator to decide when and how their work is
made public. This ensures that the creator receives recognition and credit for
their creation. The Sui generis model for AI works should be covered by
industrial copyright laws, which are supposed to protect the economic gains
made from their exploitation. In a similar line of thought, the sui generis rights
for AI-created works should not extend to moral rights.129 Since AIs do not have a
legal persona, reputation or individuality, moral rights do not apply. In its place,
this sui generis framework should deal only with economic rights and guarantee
that in this respect rightholders interested in works generated by AI creations
will be able to commercially exploit them for profitable purposes.130
Also, protection for the new sui generis rights in AI-created works should have a
limited duration of ten years or a similar period. By comparison, traditional
copyright protects works for the life of the author plus seventy additional
years.131 For many other sui generis intellectual properties, protection is
significantly shorter, and only for a duration of ten years from the date of
registration with the U.S. Copyright Office, or ten years from the date of first
commercial use anywhere in the world, whichever is earlier.132 It should also be
provided in the proposed legislation that sui generis rights in AI-created works
are protected for a ten-year duration from the date when an AI-created creation
is first made available to the public.
Sui generis rights should only extend to prevent exact copies or duplicates of AI-
generated creations.133 In traditional materials that are copyrighted,
unauthorized copying that is considerably similar to the work in copyright,
128
Ibid.
129
Benjamin Hardman and James Housel, 'A Sui Generis Approach to the Protection of AI-
Generated Works: Balancing Innovation and Authorship' (2023) SSRN
https://ssrn.com/abstract=4557004
130
Ibid.
131
17 U.S.C. § 302(a) (2023) (‘Copyright in a work created on or after January 1, 1978, subsists
from its creation and, except as provided by the following subsections, endures for a term
consisting of the life of the author and 70 years after the author’s death’).
132
Ibid 904.
133
Hardman and Housel (n 105).
34
except in cases of fair use, would amount to copyright infringement. In the case
of AI-generated creation, the threshold for infringement needs to be that of
literal copying or precise replication.134 This recognizes the fact that creation
through AI technologies is based on the training of algorithms through large data
sets, and since a non-human machine cannot be an independent author, then
lower levels of legal protection would be appropriate.
Chapter 5
5.1 Summary of Findings
This dissertation explored the intersection of AI-generated art and copyright law,
aiming to address the challenges and implications of assigning copyright to non-
human creators. The research findings indicate that current copyright laws are
insufficient to address the complexities introduced by AI technologies. A key
finding is that AI-generated works do not fit neatly within traditional copyright
frameworks, suggesting a need for legislative reform.
The theoretical implications of this study suggest a shift in how creativity and
authorship are conceptualized in the digital age. Practically, the findings highlight
the urgent need for legal clarity to ensure that both human artists and AI
developers understand their rights and obligations. The study contributes to the
ongoing debate by providing a detailed analysis of existing legal frameworks and
proposing potential pathways for adapting copyright law to accommodate AI-
generated art.
guidance, they are largely inadequate for the nuances of AI-created works. The
objectives of understanding and proposing solutions for the copyright challenges
posed by AI art were met through comprehensive legal analysis and stakeholder
interviews.
In conclusion, this dissertation highlights the pressing need for adaptive legal
frameworks that can keep pace with technological advancements. By addressing
the unique challenges posed by AI-generated art, this research contributes to a
more nuanced understanding of copyright law in the digital age. The findings
underscore the importance of ongoing dialogue between legal scholars,
technologists, and policymakers to ensure that copyright law evolves in a way
that balances innovation with protection of creators’ rights.
36
Bibliography Table:
Journal Articles:
9. Cetinic E and She J, 'Understanding and Creating Art with AI: Review and
Outlook' (2022) 18(2) ACM Transactions on Multimedia Computing,
Communications, and Applications 1
10. Fenwick M and Jurcys P, 'Originality and the Future of Copyright in an Age
of Generative AI' (2023) 51 Computer Law & Security Review 105892
12. Gervais D, 'The Machine as Author' (2020) 105(5) Iowa Law Review 2053
19. Lee E, 'Prompting Progress: Authorship in the Age of AI' (2024) 76 Florida
Law Review
22. Ng E, 'B2C2 Ltd v Quoine Pte Ltd [2019] SGHC (I) 03' (2019) Singapore
Commercial Law Review 207
26. Picht P G and Thouvenin F, 'AI and IP: Theory to Policy and Back Again –
Policy and Research Recommendations at the Intersection of Artificial
Intelligence and Intellectual Property' (2023) 54(6) IIC-International
Review of Intellectual Property and Competition Law 916
10. Lionel Bently and Brad Sherman, Intellectual Property Law (5th edn, OUP
2018).
Conference Papers:
40
Websites:
1. Basanta A, 'All We'd Ever Need is One Another' (last accessed 18 July
2024) http://allwedeverneed.com/
9. Tamim B, 'AI Image Generator Hit by $1.8 Trillion Lawsuit from Getty
Images: The Stock Image Agency Claims that Over 12 Million Copyrighted
Images Along with Their Descriptions and Metadata—Were Used to Train
Stable Diffusion, an AI Image Creator' (8 February 2023) Interesting
Engineering https://interestingengineering.com/culture/getty-images-
lawsuit-against-stability-ai
10. 'AI Will Become the New Normal': How the Art World's Technological
Boom Is Changing the Industry' (The Art Newspaper - International art
news and events, 28 February
2023) https://www.theartnewspaper.com/2023/02/28/ai-will-become-
the-new-normal-how-the-art-worlds-technological-boom-is-changing-
the-industry
11. 'As Fight over A.I. Artwork Unfolds, Judge Rejects Copyright Claim' (21
August 2023) The New York
Times https://www.nytimes.com/2023/08/21/technology/ai-artwork-
copyright.html
Cases:
1. Alfred Bell & Co v Catalda Fine Arts Inc 191 F2d 99 (2nd Cir 1951)
2. Asia Pacific Publishing Pte Ltd v Pioneers & Leaders (Publishers) Pte Ltd
[2011] SGCA 37
4. Brompton Bicycle Ltd v Chedech [2020] EWHC 104 (IPEC), [2020] FSR 25
10. Football Dataco Ltd v Yahoo! UK Ltd [2012] EWCA Civ 1759, [2013] 1 All
ER 15
13. IceTV Pty Ltd v Nine Network Australia Pty Ltd [2009] HCA 14
14. Kelley v Chicago Park District 635 F 3d 290, 304 (7th Cir 2011)
16. Time Inc. v Bernard Geis Associates 293 F Supp 130 (SDNY 1968)
17. University of London Press Ltd v University Tutorial Press Ltd [1916] 2 Ch
601
18. Urantia Foundation v Maaherra 114 F 3d 955, 964 (9th Cir 1997)
Legislation (chronologically):
Other:
3. Hardman B and Housel J, 'A Sui Generis Approach to the Protection of AI-
Generated Works: Balancing Innovation and Authorship' (2023)
SSRN https://ssrn.com/abstract=4557004