Indian Journal of Law and Legal Research Volume VI Issue III | ISSN: 2582-8878

ARTIFICIAL INTELLIGENCE (AI) AND COPYRIGHT LAW:

ANALYSIS OF ISSUES IN INTERNATIONAL IP LAWS

Adv Prashant Mali, Ph.D. - Cyber and Privacy Lawyer

Page No: 564 to 578

ABSTRACT
The use of artificial intelligence (AI) in recent years has been largely
associated with automated decision-making intended to make such process
more effective and efficient. Artificial intelligence is now capable of
generating and assisting in the production of original works of literary and
artistic nature. For example, AI recreated Dutch painter Rembrandt van Rijn’s
painting style accurately to create the Next Rembrandt which resembles the
Rembrandt original, with the use of deep learning and facial recognition
techniques.1 It is clear that while advancements in AI technology are
happening at a lightning speed, law – especially copyright law in respect of
AI-generated/assisted or Generative AI (GenAI) literary and artistic works,
needs some catching up to do. There exists a ‘grey area’ in copyright law with
regard to its application in respect of AI-generated works. At the outset, there
exists a debate on whether AI-generated works could be afforded copyright
protection in the first place, and whether they belong to the public domain or
not. There are questions over the originality of
AI-generated works, and their ownership and authorship. Copyright law’s
emphasis on ‘human authorship’ further complicates the matter, because AI
does not fall into the category of ‘person’ within the ambit of copyright law.
Hence, this paper aims to analyse the issues arising out of the application of
copyright law on AI-generated works, in the Indian as well as international
context – and suggest feasible solutions to resolve such issues.
Keywords: AI, artificial intelligence, copyright, technology
Introduction to Artificial Intelligence

Computing has travelled a long way since the days of mere number-crunching and complex
calculations. With the emergence of artificial intelligence, the boundaries of what is and
could be possible are continuously expanding at an astonishing pace. Today, AI has the
capability to write novels of such standard that the said novel even cleared the screening
round of a literature contest. Google AI has developed its own ‘child AI’, which can
outperform every other human-built AI system. The Google AI therefore has the power to
create and train its own AI. 2

1
Haochen Sun, Redesigning Copyright Protection in the era of Artificial Intelligence (2022) 107 ILR 1213,
1231
2
V.K. Ahuja Artificial Intelligence and Copyright: Issues and Challenges, ILI Law Review Winter Issue (2020)
Indian Journal of Law and Legal Research Volume VI Issue III | ISSN: 2582-8878

AI may be generally understood as “ability of machines to perform cognitive tasks like

thinking, perceiving, learning, problem-solving, and decision-making”3. The High Level
Expert Group (HLEG) appointed by the European Commission defines AI as "Artificial
intelligence (AI) systems are software (and possibly also hardware) systems designed by
humans that, given a complex goal, act in the physical or digital dimension by perceiving
their environment through data acquisition, interpreting the collected structured or
unstructured data, reasoning on the knowledge, or processing the information, derived from
this data and deciding the best action(s) to take to achieve the given goal. AI systems can
either use symbolic rules or learn a numeric model, and they can also adapt their behaviour
by analysing how the environment is affected by their previous actions”4. The World
Intellectual Property Organization (WIPO) categorizes AI into i) “expert (or knowledge-base)
systems”; (ii) “perception systems”; and (iii) “natural language systems”.5

Artificial intelligence does not refer to a single technology – it is a broad umbrella, which
encompasses various forms of processing and learning technologies. It includes machine
learning, deep learning and natural language processing. Machine learning refers to a field of
enquiry which is devoted to understanding and building methods that 'learn', that is, methods
that leverage data to improve performance on some set of tasks6. Deep learning is a class of
machine learning which performs much better on unstructured data. It utilises both
supervised as well as unsupervised algorithms. An application powered by deep learning can
learn all the relevant important features by itself, instead of relying on manual inputs by the
data scientist. Natural language processing is the systematic approach by means of a
computer that gathers knowledge on how humans use, apply, and understand language.

With respect to the works involving the use of AI, they may be classified as AI generated or
Generative AI and AI assisted works. AI generated works are generated without human
intervention. AI assisted works are, on the other hand, created with significant human
intervention.

3
Sanjivini Raina, “Artificial Intelligence through the Prism of Intellectual Property Laws” in V.K. Ahuja and
Archa Vashishtha, Intellectual Property Rights: Contemporary Developments 133-41 (Thomson Reuters, 2020).
4
High-Level Expert Group on Artificial Intelligence, “A definition of AI: Main capabilities and scientific
disciplines” December 18, 2018 available at A definition of AI (europa.eu)
5
WIPO, “WIPO Worldwide Symposium on the Intellectual Property Aspects of Artificial Intelligence”, WIPO,
March 25, 1991, available at: https://www.wipo.int/edocs/pubdocs/en/wipo_pub_698.pdf. (last visited on
January 26, 2023 ).
6
Mitchell, Tom (1997). Machine Learning. New York: McGraw Hill. ISBN 0-07-042807-7. OCLC 36417892.
Indian Journal of Law and Legal Research Volume VI Issue III | ISSN: 2582-8878

Status of copyright protection of AI/Computer-generated works across jurisdictions

None of the countries in the world grant copyright protection to AI-generated works as a
matter of right. However, there are some countries such as the UK and India where copyright
protection is guaranteed to computer-generated works. Computer-generated works could be
understood as a broader category of machine-generated works under which AI-generated
works may be included as a subset thereof. A computer-generated work refers to those
categories of works “that are created in total absence of any human intervention at the time of
the creation of the work”. 7 This is done by attributing the authorship to the person “who has
undertaken the necessary arrangements for the creation of the work”. In other jurisdictions
such as the EU and the US, insistence of human authorship, and more specifically, the
requirement to have a ‘human creator’ as an author, has prevented granting of copyright
protection to AI/computer-generated works. In China, the courts have taken a novel approach
by granting copyright protection to an AI-generated work by attributing the authorship to the
developers of the company that had developed the AI. Each of these shall be discussed
hereunder.
United Kingdom

The Copyright Designs and Patents Act, 1988 (‘CDPA’) provides copyright protection to
computer-generated works. For computer-generated works (i.e. literary, artistic, dramatic or
musical work) which is computer-generated, the author is the person by whom the
arrangements necessary for the creation of the work are undertaken”8 The author in respect of
AI-generated works could be the AI programmers or developers as ‘individuals’ or the
‘employer’, under whose direction and investment the AI programmers / developers, as
‘employees’ have undertaken ‘the necessary arrangements’. 9The term of protection for
computer-generated works is 70 years from the date of creation10. In Nova Productions v
Mazooma Games and Others, the court recognized the authorship of the programmer /
developer of computer-generated work by taking into account that “he [the programmer]
devised the appearance of the various elements of the game and the rules and logic by which
11
each frame is generated and he wrote the relevant computer program”. . This reflects a
positive approach towards granting copyright protection to computer-generated

7
Perry, Mark and Margoni, Thomas, "From Music Tracks to Google Maps: Who Owns Computer-generated
Works?" (2010). Law Publications. Paper 27. http://ir.lib.uwo.ca/lawpub/27
8
Copyright Designs and Patents Act 1988, s.9 (3)
9
Paul Lambert, Computer Generated Works and Copyright: Selfies, Traps, Robots, AI and Machine Learning,
European Intellectual Property Review
10
Copyright Designs and Patents Act 188 s.12 (3)
11
Nova Productions Ltd v Mazooma Games Ltd[2006] EWHC 24 (Ch) (20 January 2006).
 Indian Journal of Law and Legal Research Volume VI Issue III | ISSN: 2582-8878

works.Though, the UK approach has attracted criticism as well because copyright protection
to computer-generated works merely ‘identifies the contributor for computer-generated works
but did not address works which were created in absence of any human contribution”12
India

In India, the Copyright Act of 1957 grants copyright protection to computer-generated works.
The term ‘author’ in the Copyright Act, in respect of “any literary, dramatic, musical or
artistic work which is computer- generated, is “the person who causes the work to be
created”13. Though copyright protection is extended to computer-generated works by
considering the persons who have caused the work to be created as the ‘author’ - yet the
personhood of the author is assumed by default to be ‘natural personhood’. The Copyright
Act, 1957 provides for authorship to legal / artificial persons in respect of certain classes /
categories of ‘work’14. However, in respect of computer-generated work, no such express
mention of ‘legal / artificial person’ is present. Hence, the possibility of a person “who has
caused the work to be created”, to be a ‘an artificial person’ or a ‘legal person’ is implicitly
excluded.15

United States

In the United States of America (the U.S), the Copyright Act, 1976, unlike equivalent
legislations in the UK and India, fails to address the question of authorship in respect of
computer-generated works, and hence, the legislation, prima facie, does not appear to grant
copyright protection to computer-generated works. However, a perusal of US case law will
reveal that if the ‘computer’ (and by that extension, the AI), is shown to be a ‘tool’ or a
‘medium’ in the process of creating the work, then such work is granted copyright protection
as per law16 and precedent for the same exists in the United States, as exemplified in the case
of Burrow-Giles Lithographic Co. v. Sarony, wherein it was held that copyright of the author
vested in a photograph because the machine (the camera in this case) was a tool in the process
in creating “an original work of art.”17 However, in the case of creative work generated /
created autonomously by AI, copyright protection is not granted unless human ingenuity and
influence is not demonstrated. 18 The U.S. Copyright Office categorically states

12
Haochen Sun supra note 1
13
Copyright Act 1957, s.2 (d) (vi)
14
Copyright Act 1957, s.17
15
Chakraborty, Avishek, Authorship of AI Generated Works under the Copyright Act, 1957: An Analytical
Study (July 26, 2019). Nirma University Law Journal: Volume-8, Issue-2, July-2019 , Available at SSRN:
https://ssrn.com/abstract=3443006
16
17 U.S.C. § 102(a).
17
Burrow-Giles Lithographic Co. v. Sarony, 111 U.S. 53 (1884).
18
U.S. Copyright Office, Compendium of U.S. Copyright Office Practices § 306 (3rd ed. 2014).
Indian Journal of Law and Legal Research Volume VI Issue III | ISSN: 2582-8878

that “[the office] will not register works produced by a machine or mere mechanical process
that operates randomly or automatically without any creative input or intervention from a
human author.”19 This has resulted in computer-generated works autonomously created
without substantial human involvement to fall in the public domain. A landmark case in this
regard is Naruto v. Slater. In Naruto v. Slater, a photographer named David Slater, left his
camera unattended in the wild which resulted in a photograph clicked by a monkey. Slate
claimed that the copyright in the said photo vested in him since it was his camera, however, it
was argued before Ninth Circuit Court of Appeals that the copyright, in fact, belonged to the
monkey dubbed ‘Naruto’. The court held that since ‘animals’ were not persons within the
meaning as envisaged in the U.S. Copyright Act, 1976, hence, copyright vested neither in the
photographer nor the monkey, and hence, the photo was consigned to public domain20.
Though not directly related to computer or AI generated works, the judgement in the above
case reveals the firmly anthropocentric interpretation of copyright law in the United States.
European Union
A similar anthropocentric interpretation of copyright law prevails in the European Union -
which results in failure to provide copyright protection to AI generated works falling within
the larger umbrella of computer-generated works. The Court of Justice of the European
Union (CJEU) laid down the prevailing standard of ‘originality’ (required for a work to
qualify for copyright protection) in the Infopaq judgement wherein it was held that “
computer programs, databases or photographs are protected by copyright only if they are
original in the sense that they are their author’s own intellectual creation.”.21 EU laws, such as
the Copyright Term Directive and the body of case laws developed by the CJEU reflect a
preference for the personality theory of IP rights. The Copyright Term Directive, in respect of
photographic works for example, says that such work “is to be considered original if it is the
author’s own intellectual creation reflecting his personality, no other criteria such as merit or
purpose being taken into account” 22In the Dataco judgement it was held that “[The] criterion
of originality is satisfied when, through the selection or arrangement of the data which it
contains, its author expresses his creative ability in an original manner by making free and
creative choices . . . and thus stamps his ‘personal touch.’”23. Hence, the position of copyright
law in the EU especially with regard to computer-generated works is that the requirement of
19
Id. at § 313.2
20
Naruto v. Slater 888 F.3d 418
21
Infopaq International v. Danske Dagblades Forening [2009] ECDR 16 (Case C-5/08).
22
Directive 2006/116, of the European Parliament and of the Council of 12 December 2006 on the Term of
Protection of Copyright and Certain Related Rights, art. 6, 2006 O.J. (L 372) 12, 14 (EC)
23
Football Dataco Ltd and others v. Yahoo! UK Ltd and others, 1 March 2012 (Case C-604/10).
Indian Journal of Law and Legal Research Volume VI Issue III | ISSN: 2582-8878

the creation to be the author’s own and the reflection of the author’s personal touch (i.e.
his/her creativity) is paramount for any work to be accorded copyright protection.
However, with the advent and proliferation of AI generated works, EU legislators are
rethinking upon the traditional paradigm of copyright law. The EU Parliamentary Committee
on Legal Affairs, for example, has noted that “At a time when artistic creation by AI is
becoming more common, we seem to be moving towards an acknowledgement that an AI-
generated creation could be deemed to constitute a work of art on the basis of the creative
result rather than the creative process.”.24 EU lawmakers have begun recognising that
imposition of the human author requirement may impede the protection of copyrightable
computer-generated works25, and that computer-generated works deserve the same level of
protection as traditional works because both play a key role in expanding cultural heritage.26
The present compromise version of the EU AI Act requires generative AI providers to
publicly publish details of copyright works used in training in an effort to raise awareness of
generative AI and copyright violation. A sufficiently complete explanation of the usage of
training data protected by copyright law must be "documented and made publicly available"
by the "provider" of a foundation model incorporated into or specialised into a generative
artificial intelligence system, according to Article 28b(4)(c). Prior to the model being
marketed or used in an EU service, this needs to be completed (Article 28b(1)). An argument
that AI implementers should reduce the risks of third-party IP infringement under broad
standards for "governance" can be made by combining wording from the EU AI Act's
recitals. "Governance" is associated with "mitigating risks to fundamental rights," which
include "intellectual property rights" (Recital 28a), according to Recital 58a. These
copyright-related obligations aren't, however, specifically stated in the articles. Recital 58a
links ‘governance’ to ‘mitigating risks to fundamental rights’ and such rights include
‘intellectual property rights’ (Recital 28a). However, no such obligation related to copyright
is set out expressly in the articles.

24
EUROPEAN PARLIAMENT COMM. ON LEGAL AFFS., REPORT ON INTELLECTUAL PROPERTY
RIGHTS FOR THE DEVELOPMENT OF ARTIFICIAL INTELLIGENCE TECHNOLOGIES 9 (Oct. 2, 2020),
https://www.europarl.europa.eu/doceo/document/A-9-2020-0176_EN.pdf [https://perma.cc/ D9AW-SMGJ].
25
Madeleine de Cock Buning, Artificial Intelligence and the Creative Industry: New Challenges for the EU
Paradigm for Art and Technology by Autonomous Creation, in RESEARCH HANDBOOK ON THE LAW OF
ARTIFICIAL INTELLIGENCE 511, 533 (Woodrow Barfield & Ugo Pagallo eds., 2018)
 Indian Journal of Law and Legal Research Volume VI Issue III | ISSN: 2582-8878

China
In recent years Chinese courts have had the opportunity to engage with copyright-related
disputes pertaining to AI generated works, leading to radical departures from the traditional,
anthropocentric view of copyright law in relation to computer-generated works, and in
particular, AI-generated works. In Tencent v Shanghai Yingxun Technology Co. Ltd
(hereinafter referred to as the ‘Dreamwriter’ case), the Nanshan District People’s Court held
that an article, generated by Tencent’s AI software, the Dreamwriter, was protected by
copyright and authorship thereof was attributed to the software developers of the AI in
question, i.e. Tencent. The judgement delivered in the Dreamwriter case is interesting from a
comparative law perspective because it is rarely that the court finds that skill and judgement
of the AI developers was involved in the creation of an AI-generated work.27 The
Dreamwriter software, in order to generate the article, collected data from verified sources,
compiled it, verified the data and then generated the article. The court found that this would
not have been possible without “established rules, algorithms, and templates” which were
personally selected by the developers28. This included other aspects such as “arrangement of
data type, data format, setting of the corpus and training of the intelligent verification
algorithm model”, none of which would have been possible without human contribution.29
Hence, the court found that decisions and choices made by developers in respect of the AI
software that they had developed, were sufficient evidences of the ‘originality’ of the AI
generated work, because the AI could not have possibly created the article without the
specific set of instructions, algorithms designed, data fed and other technical aspects included
by the AI developers. In effect, this position of law evolved in China is similar to that
existing in the UK. However, in China, the court went ahead beyond the mere identification
of the correct author in respect of the computer-generated work, and actually found evidence
of originality in the work in the form of the contribution put in by developers of the said AI.

26
Haochen Sun supra note 1
27
Rita Matulionyte and Jyh-An Lee, "Copyright in AI-generated works: Lessons from recent developments in
patent law" (2022) 19:1 SCRIPTed 5 https://script-ed.org/?p=4036
DOI: 10.2966/scrip.190122.5
28
Id.
Indian Journal of Law and Legal Research Volume VI Issue III | ISSN: 2582-8878

Arguments in favour of granting Copyright Protection to AI-generated works

Proponents in favour of granting copyright protection in AI-generated works have argued that
lack of copyright protection to AI-generated works will result in disincentivization of
technological advancements which runs counter to the aim of copyright, if not, IP law -
which is to incentivise innovation. Lack of copyright protection could reduce the investment
in AI30. Plus, copyright protection would incentivise copyright holders in the AI-generated
work to exercise better control over the use of AI-generated work - for example, preventing
its use for the purposes of copyright infringement.31

If copyright protection were to be granted to AI generated works, then who would be the
author of such work? Proponents in favour of granting copyright protection to AI generated
works differ on the issue of authorship - but generally, three key entities have been argued to
be eligible for attribution of authorship of the AI-generated work. These are: (i) The AI itself;
(ii) The AI developer / programmer; (iii) The user.

Those in favour of attributing authorship to the AI itself argue that copyright law must be in tandem
with the developments in technology. As such, it is argued that copyright law should be cognizant of
the new reality where concepts such as digital authorship will emerge and “digital works (i.e., software
programs) will, relatively autonomously, produce other works that are indistinguishable from works of
human authorship”32. Indeed, the European Union is now proposing the grant of electronic personhood
to autonomous robots33. In Europe, an example has already been set - the Artificial Intelligence Virtual
Artist (AIVA) has been granted the status of ‘composer’ (i.e. a music composer) by SACEM
(Société des auteurs, compositeurs et éditeurs de musique) in France.34

29
Id.
30
Tim W. Dornis, Artificial Creativity: Emergent Works and the Void in Current Copyright Doctrine, 22 YALE
J.L. & TECH. 1, 20 (2020)
31
Shlomit Yanisky-Ravid, Generating Rembrandt: Artificial Intelligence, Copyright, and Accountability in the
3A Era—The Human-Like Authors are Already Here—A New Model, 2017 MICH. ST. L. REV. 659, 703-7 04
32
Annemarie Bridy, Coding Creativity: Copyright and the Artificially Intelligent Author, 2012 STAN. TECH.
L. REV., 1, 24–27
33
V.K. Ahuja supra note 2
Indian Journal of Law and Legal Research Volume VI Issue III | ISSN: 2582-8878

Those in favour of attributing AI authorship to the AI programmer / developer argue that

precedent for the same already exists - most notably in the CDPA Act of the UK, 1988 as
discussed earlier - under which, the authorship for computer-generated works is attributed to
the person who undertakes the ‘necessary arrangements’ for the creation of that work.

Though in the case of AI-generated works, the work is generated by the AI with minimal
human input, yet the contribution of the programmers / developers in the development stage
of the AI responsible for the generation of that work cannot be overlooked in the process of
creation of the AI generated work35. The Chinese court in the Dreamwriter case echoed a
similar position. Hence, it is argued that copyright protection should be provided to AI
generated works, and the authorship thereof should be attributed to the developers /
programmers. Those in favour of end-users being attributed as the authors of the AI-
generated work argue that attributing authorship to AI programmers / developers would lead
to ‘excessive protection’ or ‘double protection’ guaranteed to programmers / developers,
because the programmers / developers are already the copyright holders of the source code
of the AI itself36. Plus, it has been argued that the programmer may not be aware of all the
creative works that may be generated by the AI. It is the user, in the case of publicly available
AI, that makes the selection of data and initiates the process of creating the work. Hence, it
has been argued on these grounds that it is the user who should be the author of the AI
generated work.37 To summarise, those in favour of attributing human authorship to AI-
generated works, and thereby, grant copyright protection to AI generated works envision an
enlarged concept of authorship “beyond persons who actually create copyrightable
expression to persons who originate the process of creating copyrightable expression.” 38

Arguments in favour of granting Copyright Protection to AI-generated works

The arguments against the attribution of authorship of AI-generated works to the AI itself
stem from the belief that copyright law envisages authorship as purely human. Though

34
Ed Lauder, “Aiva is the first AI to officially be Recognised as a Composer”, AI Business, October 3, 2017,
available at: https://aibusiness.com/document.asp?doc_id=760181 (last visited on January 21, 2023)
35
Haochen Sun, supra note
36
Robert C Denicola, ‘Ex Machima: Copyright Protection for Computer-Generated Works’ (2016) 69 Rutgers U
L Rev 251, 280-281.
37
Sik Cheng Peng, “Artificial Intelligence and Copyright: The Author’s Conundrum”, WIPO-WTO Colloquium
Papers,181 ( 2018)
38
Robert C Denicola supra note 36 at 283
Indian Journal of Law and Legal Research Volume VI Issue III | ISSN: 2582-8878

‘author' has not been specifically defined as ‘human’ in any IP rights instrument - for
example the Berne Convention, yet it has been traditionally interpreted and understood that
the term ‘author’ refers only and only to ‘human author’.39 Berne Convention is argued to be
anthropocentric - “[p]eople, rather than machines, have always been the object of the [Berne]
Convention, and, from the point of view of principle, doctrine and practicality, this object
should continue to be upheld”.40
The labour and personality theories of IP support the contention that the author of an original
work can be considered as a human only and not any other entity. The labour theory states
that for the authorship of any work, the contribution of human labour to create an original
work is central to determining whether any right of a person or person (s) vests in that work.
41
The author has the right to enjoy the fruits of the work produced by his/her labour. The
personality theory states that a work is the manifestation of the human will and therefore an
extension of his/her personality.42 Authorship is connected with ‘humanness’ and “human
communication is the very point of authorship as a social practice—indeed, as a condition of
life”. 43 With specific reference to art, it has been argued that “[c]reating a work of art is one
of those activities we reserve for humans” because it is an act of self-expression.44 Hence,
there has to be a causal link between the output generated and the human author - otherwise
such work would fall outside the domain of copyright law.45
Some of the practical issues that may arise owing to AI authorship include determination of
liability in the event of an AI-generated work violating any law - such as those against
defamation or obscenity. In the event of an AI-generated work infringing copyright in another
work, how liability of the AI could be determined? These are questions yet to be answered,
and hence, present a roadblock to the idea of AI authorship.46
Attribution of human authorship to AI generated works has also its fair share of criticism. For
example, the idea of attributing authorship to end users is criticised on the grounds that end

39
Sam Ricketson and Jane C. Ginsburg, International Copyright and Neighbouring Rights: The Berne
Convention and Beyond, Vol. 1, 2nd edn. (Oxford University Press, 2005), 358.
40
Sam Ricketson, “The 1992 Horace S. Manges Lecture – People or Machines: The Berne Convention and the
Changing Concept of Authorship” Columbia-VLA Journal of Law & the Arts 3, (1991).
41
William Fisher, Theories of Intellectual Property, in NEW ESSAYS IN THE LEGAL AND POLITICAL
THEORY OF PROPERTY 168, 171 (Stephen R. Munzer ed., 2001).
42
Id.
43
Craig & Ian Kerr, The Death of the AI Author, 52 OTTAWA L. REV. 31, 67 (2021)
44
Lev Grossman, “2045: The Year Man Becomes Immortal”, Time Magazine, Published February 10, 2011.
Accessed January 14, 2023. http://content.time.com/time/magazine/article/0,9171,2048299,00.html
45
Mezei, Péter, 'You AIn’t Seen Nothing yet' – Arguments against the Protectability of AI-generated Outputs by
Copyright Law (July 20, 2021). Available at SSRN: https://ssrn.com/abstract=3890051or
http://dx.doi.org/10.2139/ssrn.3890051
46
V.K. Ahuja, supra note 2
Indian Journal of Law and Legal Research Volume VI Issue III | ISSN: 2582-8878

users have the least contribution to the development of AI and hence authorship should not be
granted to them.47
Furthermore, proponents of AI authorship tend to focus on the creativity of the output
generated to justify granting authorship to a non-human entity such as AI. Opponents of AI
authorship however note that creativity is not the prerequisite to determining originality in a
vast majority of the jurisdictions. Originality “is generally fixed to authorship and subject
matter, both of which are closely connected to humans and human achievements”48. Hence it
is on the foregoing grounds that AI authorship is opposed and criticised.

AI-generated works and Copyright Protection: The Way Forward

By all accounts, the future of AI is on an upward trajectory. In the future, AI would not only
be capable of creating original works of literary, artistic or dramatic nature, but also replace
humans in jobs traditionally considered the sole purview of human beings. In such a scenario,
it is important, according to the proponents of copyright protection to AI-generated works,
that a definite roadmap be chalked for the regulation and protection of AI-generated works.
And for that, copyright law needs significant re-interpretation and re-design. Currently, this
can be done through two ways: (i) Establishing a sui-generis legislative framework for
AI-generated works and (ii) Applying the work-for-hire doctrine to ensure copyright
protection to AI-generated works.
Yet, there are those who believe that the future of AI-generated works is one, where, by virtue
of their autonomous creation, they should not receive copyright protection, thereby upholding
the fundamental anthropocentricity of copyright law. Advocates of such viewpoints hold that
AI-generated works should be consigned to the public domain.
All of the foregoing perspectives shall be discussed hereunder.

Sui-generis framework for AI-generated works

A sui-generis framework means protecting AI-generated works as a separate class of works.

Such a framework of protection would be independent from copyright and other IP rights. It
is argued that establishing a sui-generis framework would solve many problems stemming
from the lack of compatibility of AI-generated works with copyright law. For example, as per
U.S. copyright law, the work needs to be fixed on a tangible medium of expression, because it
is then that the author could exercise control on the final form of that work. In the case of
AI-generated works, the AI software / system uses its own intelligence to create a work in a
form that may or may not be predicted or intended by the human who is using the AI or who

47
Kalin Hristov, ‘Artificial Intelligence and the Copyright Dilemma’ (2017) 57 IDEA 431, 445
48
Id.
Indian Journal of Law and Legal Research Volume VI Issue III | ISSN: 2582-8878

has developed the AI49. Hence, the “causal link between the creativity of the human creator
and the resulting work is weakened or even broken”. 50 Another reason for having sui-generis
protection is that AI cannot be simply viewed as a ‘tool’ used in the creation of literary,
artistic or dramatic works. This distinguishes AI-generated works from other machine-
generated works such as photographs. In countries such as the US and the UK, machine-
generated works have been granted copyright protection on the premise that machines are
‘tools’ used in the creation of the work51. The ‘tools’ inherently lack creativity, and it is
instead the human who employs his creativity by using the machine as tools to produce the
creative work.52 In the case of AI-generated works, AI has already matched or even
surpassed human intelligence. AI autonomously creates literary, artistic or dramatic works
and its decision making process is independent to the extent that even the programmers/
developers have no idea how the AI-decision making process works (this is known as the
black-box problem in AI).53 Hence, it is due to the foregoing reasons that a separate sui-
generis framework for AI-generated works is required. An example of how this would look
like could be found in the protection regime for databases in the EU. The EU has enacted a
separate sui-generis framework for the protection of databases vide the Database Directive.
The Database Directive “protects the contents of a database and operates irrespective of
whether the database or any of its content is copyrighted”.54 The term of protection is for 15
years.

Work-for-Hire doctrine

The work-made-for-hire doctrine is proposed mainly by American legal scholars and as such,
is a U.S.-specific solution. However, other jurisdictions may still take cue from this approach.
The work-made-for-hire doctrine stipulates that if the work has been created by an employee
in the course of employment, then the employer shall be deemed to be the author of the work.
55
The employer may be a natural person or a juristic person like a company. It is proposed
that the ‘employer’ in the case of AI-generated works, could be considered as the
programmers or the developers of the AI, or the company that developed the AI. The

49
Haochen Sun supra note 1
50
Daniel J. Gervais, The Machine as Author, 105 IOWA L. REV. 2053, 2072 (2020)
51
See Burrow-Giles Lithographic Co. v. Sarony, 111 U.S. 53 (1884) in the U.S. and Express Newspapers plc v.
Liverpool Daily Post & Echo [1985] FSR 306 in the U,K
52
Gervais supra note 50
53
Samantha Fink Hedrick, I “Think,” Therefore I Create: Claiming Copyright in the Outputs of Algorithms, 8
N.Y.U. J. INTELL. PROP. & ENT. L. 324, 371 (2019).
54
Council Directive 96/9, of the European Parliament and of the Council of 11 March 1996 on the Legal
Protection of Databases, ch. 2–3, 1996 O.J. (L 77) 25–27 (EC).
55
U.S. Copyright Office, Circular 9: Works Made for Hire (Sep. 2012),
https://www.copyright.gov/circs/circ09.pdf [https://perma.cc/V86PSA8A]
Indian Journal of Law and Legal Research Volume VI Issue III | ISSN: 2582-8878

‘employee’ in the case of AI-generated works could be considered as the AI machine /

software / system.56 This would lead to two outcomes - the AI system would be granted
authorship by the virtue of the fact that it autonomously created the work in question, and
secondly, the developer or owner of the AI could be “considered the author for the purpose of
the title.”57
This doctrine may be applied in the Indian context as well. S.17 of the Copyright Act, 1957
states the author shall be the first owner of the copyright, provided that “in the case of a work
made in the course of the author's employment under a contract of service or apprenticeship,
to which clause (a) or clause (b) does not apply, the employer shall, in the absence of any
agreement to the contrary, be the first owner of the copyright therein”58 Assuming that the AI
machine / software / system is the author of the AI-generated work, the first owner of the
work would be the programmer / developer of the AI since, as per the re-interpreted work-
made-for-hire doctrine, the programmers / developers of the AI would be the ‘employers’.
Hence, these are some of the key approaches by which AI-generated works could be granted
copyright protection.
The Case for Public Domain

Those in favour of assigning AI-generated works to public domain refer specifically to those
categories of works which are produced by ‘autonomous AI’ technologies. These AI
technologies require almost no human input and can create new, original works by utilising
deep learning. An autonomous AI system is able to “independently sense and select among
different courses of action” and then make decisions and “accomplish goals based on its
knowledge and understanding.”59.
It is argued that since autonomous AI technologies do not require incentive to create works,
hence, it does not require copyright protection. One of the key motives of granting copyright
protection to any literary, artistic or dramatic work is to incentivise the creation of original
works, and innovation in general. However, since autonomous AI technologies can create
such works anytime when prompted, they do not require incentive, and as such, should be

56
Kalin Hristov supra note 47
57
Elvia Arcelia Quintana Adriano, The Natural Person, Legal Entity or Juridical Person and Juridical
Personality, 4 PENN. ST. J. L. & INT'L AFF. 363, 366 (2015)
58
Copyright Act 1957, s.17 (c)
59
Will Knight, The Dark Secret at the Heart of AI, MIT TECH. REV. (Apr. 11, 2017), https://
www.technologyreview.com/s/604087/the-dark-secret-at-the-heart-of-ai [https://perma.cc/TE86- Z776];
Indian Journal of Law and Legal Research Volume VI Issue III | ISSN: 2582-8878

consigned to the public domain60. Another key reason is that autonomous AI agencies lack
human agency61. It is settled law in copyright law that authorship vests in natural persons
only. AI technologies cannot be deemed as persons within the ambit of copyright law because
they don’t have the capacity to sue or to be sued62. Also, since no cost is incurred in the
creation of works (by the AI), and multiple copies of the same work could be generated by
the AI, it is but natural that these works should be publicly accessible for free. 63

Conclusion

The future of AI is exciting, yet fraught with uncertainties galore. It promises to revolutionise
all spheres of human life, yet it upends many fundamental notions and conceptions. With
respect to AI-generated works, the fact that such works may call into question the causal link
between the human author and the AI-generated work has lead to a substantial re-think on the
inherent anthropocentricity of copyright law. Since the inception of IP rights, the gamut of IP
laws, including the copyright law, has adapted itself to the rapid changes in technology and
the challenges associated therewith (for example, computer programs, photographs etc.). The
challenge posed by AI-generated works is also a surmountable one, but it would require
abandoning some of the central tenets of copyright law and incorporating radical new
concepts to cover the unpredictable possibilities that AI would bring along. IP rights
conventions / treaties worldwide such as the TRIPS and Berne Convention would have to
come to terms with ‘digital authorship’ and ‘recognition of AI (s) as legal persons’. So will
the domestic legislations. Already, the stage has been set with the laws prevailing in the UK
and India as discussed in the foregoing sections of this paper. The time is ripe for an
expansion of the definition of computer-generated works to include AI-generated works
specifically, and provide protection to such works either by re-interpreting existing doctrines
of copyright law or establishing sui-generis frameworks. Either ways, it is the need of the
hour that the grey area created with the advent of AI-generated or Generative AI works be
resolved so that copyright law is 21st century ready. Man chose AI as his own destination so
why deny its rights which now can be done by limitations or the AI would be judge and jury
and choose its own copyright right and sue the mankind coz we were indolent.

60
Ayush Pokhriyal and Vasu Gupta, “Artificial Intelligence Generated Works under Copyright Law”, 6(2) NLUJ
Law Review 116 (2020)
61
Haochen Sun supra note 1
62
V.K. Ahuja supra note 2
63
V.K. Ahuja supra note 2