DATE DOWNLOADED: Mon Jan 22 01:20:45 2024

SOURCE: Content Downloaded from HeinOnline

Citations:
Please note: citations are provided as a general guideline. Users should consult their preferred
citation format's style manual for proper citation formatting.

Bluebook 21st ed.

Patrick Zurth, Artificial Creativity? A Case against Copyright Protection for
AI-Generated Works, 25 UCLA J.L. & TECH. i (2021).

ALWD 7th ed.

Patrick Zurth, Artificial Creativity? A Case against Copyright Protection for
AI-Generated Works, 25 UCLA J.L. & Tech. i (2021).

APA 7th ed.

Zurth, Patrick. (2021). Artificial creativity? case against copyright protection for
ai-generated works. UCLA Journal of Law and Technology, 25(2), i-18.

Chicago 17th ed.

Patrick Zurth, "Artificial Creativity? A Case against Copyright Protection for
AI-Generated Works," UCLA Journal of Law and Technology 25, no. 2 (Spring 2021): i-18

McGill Guide 9th ed.

Patrick Zurth, "Artificial Creativity? A Case against Copyright Protection for
AI-Generated Works" (2021) 25:2 UCLA JL & Tech i.

AGLC 4th ed.

Patrick Zurth, 'Artificial Creativity? A Case against Copyright Protection for
AI-Generated Works' (2021) 25(2) UCLA Journal of Law and Technology i

MLA 9th ed.

Zurth, Patrick. "Artificial Creativity? A Case against Copyright Protection for
AI-Generated Works." UCLA Journal of Law and Technology, vol. 25, no. 2, Spring 2021,
pp. i-18. HeinOnline.

OSCOLA 4th ed.

Patrick Zurth, 'Artificial Creativity? A Case against Copyright Protection for
AI-Generated Works' (2021) 25 UCLA JL & Tech i Please note:
citations are provided as a general guideline. Users should consult their preferred
citation format's style manual for proper citation formatting.

-- Your use of this HeinOnline PDF indicates your acceptance of HeinOnline's Terms and
Conditions of the license agreement available at
https://heinonline.org/HOL/License
-- The search text of this PDF is generated from uncorrected OCR text.
-- To obtain permission to use this article beyond the scope of your license, please use:
Copyright Information
Spring 2021 www.uclajolt.com Volume 25, Issue 2

UCLA Journal of Law & Technology

ARTIFICIAL CREATIVITY?

A CASE AGAINST COPYRIGHT PROTECTION FOR AI-GENERATED WORKS

Patrick Zurtht

Abstract

Today, many areas of our daily lives are determined by artificial intelligence (AI). Machines
program software, translate texts rapidly, create beautiful images, and design fashion efficiently.
They are capable of superhuman performances. Furthermore, machines make the impression of
boundless creativity. AI's achievements in traditional areas of copyright subject matters
inevitably raise the question of legal protection through an exclusive right. This Article begins
by exploring Al technology's various accomplishments (Introduction). Then, in Part I, it
outlines the legal status quo with respect to those developments. And Part II discusses potential
copyright protection, mainly from the perspective of creativity. Finally, Part III, IV, and V
address other arguments regarding copyright protection for Al products, such as the economic
aspects of incentive and market failure. This Article contemplates the issue from an international
perspective and concludes that neither copyright nor other similar protection rights, i.e. sui
generis rights (which already exist for different subject matters, for example, in Europe), should
be implemented. This finding holds true regardless of the legal jurisdiction, may it be common
law or civil law.

T Dr. iur., LL.M. (Stanford). Postdoctoral Fellow at the Chair for Private Law and Intellectual Property Law with
Information- and IT-Law (GRUR-chair) (Prof. Dr. Matthias Leistner, LL.M. (Cambridge)) at Ludwig Maximilian
University of Munich, Germany. Thanks to Professor Leistner, Abhilasha Vij, Sandra Stadler and the JOLT
Editorial Team for helpful comments.
INTRODUCTION: Al's IMPORTANCE AND ACHIEVEMENTS ................................................................. 1

I. ANTHROPOCENTRIC COPYRIGHT ............................................................................................... 3

II. CREATIVITY AS COPYRIGHT' S BASIS.........................................................................................8

A. Outlining Creativity ............................................................................................................ 10

B . A I's M ira g e ......................................................................................................................... 1 0

III. ECONOMIC ASPECTS OF EXCLUSIVE PROTECTION...................................................................13

A. Incentives and Market Failure ............................................................................................ 13

B. Skepticism Tow ards M onopolies........................................................................................15

IV. SUPERSESSION OF HUMAN CREATION ..................................................................................... 15

V. RECOGNIZABILITY AND TRANSPARENCY ................................................................................. 16

CONCLUSION...................................................................................................................................17

ii
 Artificial Creativity?
A Case Against Copyright Protection for AI-
Generated Works
PatrickZurth

Introduction: AI's Importance and Achievements

The term "Artificial Intelligence" (AI) encompasses a broad scope of technologies. 1 "Machine
learning" often falls under Al and refers to self-learning algorithms. Machine learning
technology learns from a large number of example cases and abstracts a general rule from them.
After a learning phase, these findings can then be applied to other cases. Thus, machine learning
software can independently and continuously improve its processes by observing and evaluating
input by people or occurrences or by self-observation. Self-learning algorithms have been in
place and utilized since the 1960s. At that time, there existed already a painting machine and a
translation robot used by the American Air Force.2 However, there is now much more
computing power available, allowing these algorithms to improve.3 In addition to machine
learning technology, there are also artificial neural networks (ANN), which are modeled after
natural neural networks and sometimes achieve even better results than machine learning.

The tremendous computing power explains the rapid increase in AI's relevance. Nowadays,
intelligent machines play a major role in many areas of our daily lives. For example, Al
machines program software, translate texts, create images, compose music, and design fashion,
in addition to performing a myriad of other tasks ranging in difficulty and complexity. As a
result, significant economic interests are at stake. Worldwide, AI-generated revenue is estimated

1See Shlomit Yanisky-Ravid, GeneratingRembrandt: Artificial Intelligence, Copyright, andAccountability in the

3A Era The Human-Like Authors Are Already Here A New Model, 2017 MICH. ST. L. REv. 659, 673-76 (2017)
(discussing the difficulty of finding a definition and outlining the possible approaches to defining "Artificial
Intelligence"); Atilla Kasap, Copyright and Creative Artificial Intelligence (Al) Systems: A Twenty-First Century
Approach to Authorship ofAI-Generated Works in the United States, 19 WAKE FOREST J. BUs. & INTELL. PROP. L.
335, 339-42 (2019).
2 Friedrich Karl Fromm, DerApparat als geistiger Schopfer, GEWERBLICHER RECHTSSCHUTZ UND URHEBERRECHT

304, 304-05 (1964) (Ger.).

3 See Ryan Calo, Artificial Intelligence Policy: A Primerand Roadmap, 51 U.C. DAvIS L. REv. 399, 402 (2017)
("[A] vast increase in computational power and access to training data has led to practical breakthroughs in machine
learning .... "); Daryl Lim, Al & IP: Innovation & Creativity in an Age ofAccelerated Change, 52 AKRON L. REv.
813, 827 (2018) ("[C]omputer processors attained the miniaturization, sophistication, and power needed for Al to
take off.").

1
to grow from nearly $8 billion in 2016 to more than $47 billion in 2020.4 And by 2025, AI-
generated revenue is estimated to be up to $126 billion.5

This Article cannot expound all of AI's various achievements. Three examples of AI's success
in typically copyright-related areas, however, shall be singled out and discussed for the purposes
of this Article.

In a project called "You Can't Know My Mind" in 2013, an Al software painted portraits of
people, whereby it employed styles depending on its "mood." 6 That "mood" depended on
whether the reports in the local daily newspapers were positive or negative. Once, the program
even refused to paint at all because of its negative "sentiments."

An even more famous example is the project "The Next Rembrandt"-a collaboration of
different institutions that created a painting in Rembrandt's style.' Experts from various fields
collected data by examining the works of Rembrandt for characteristics and compiled them in a
database. The data were collected by means of 3D scans and face recognition software.
Subsequently, they determined the subject of the new picture.' With this profile, the developers
began to extract characteristics from all images that matched the profile. With their findings on a
typical "Rembrandt face," they created a new painting using a 3D printer. The final work
consists of 148 million pixels and is based on 168,263 fragments of Rembrandt's paintings.

In 2016, SONY CSL Research Lab developed Flow Machines, an Al system that composes new
tunes by exploiting a huge database of songs.9 This program created, for instance, the melody of
"Daddy's Car" in the style of The Beatles. 10

4 KAY FIRTH-BUTTERFIELD & YOON CHAE, WORLD ECON. F. CTR. FOR THE FOURTH INDUS. REVOLUTION,
ARTIFICIAL INTELLIGENCE COLLIDES WITH PATENT LAW 5 (2018),
http://www3.weforum.org/docs/WEF_48540_WPEndofInnovationProtectingPatentLaw.pdf.
5 Shanhong Liu, Revenues From the Artificial Intelligence (Al) Software Market Worldwide From 2018 to 2025,
STATISTA (Nov. 2020), https://www.statista.com/statistics/607716/worldwide-artificial-intelligence-market-
revenues/.
6 Madeleine de Cock Buning, Autonomous Intelligent Systems as CreativeAgents Under the EU Frameworkfor
Intellectual Property, 7 EUR. J. RISK REG. 310, 313 (2016). For more information on that project, see also Simon
Colton, The PaintingFool, http://www.thepaintingfool.com/galleries/youcantknowmymind/ (last visited Nov.
29, 2020).
? ING, The Next Rembrandt, THE NEXT REMBRANDT, https://www.nextrembrandt.com/ (last visited Nov. 29, 2020);
see also The Next Rembrandt, The Next Rembrandt, YOUTUBE (Apr. 5, 2016),
https://www.youtube.com/watch?v=IuygOYZINgo; Mark Brown, 'New Rembrandt'to Be Unveiled in Amsterdam,
GUARDIAN (Apr. 5, 2016), https://www.theguardian.com/artanddesign/2016/apr/05/new-rembrandt-to-be-unveiled-
in-amsterdam; Yanisky-Ravid, supra note 1, at 663-64.
8 A high percentage of Rembrandt's opus were portraits. The Next Rembrandt, supra note 7. The most common

matches revealed the following characteristics for the subject of the picture: portrait of a Caucasian man with facial
hair, age 30 to 40 years, dark clothes with collar, hat-wearer, head turned to the right. Id.
9 See Sony CSL, Daddy's Car: A Song Composed by Artificial Intelligence In the Style of the Beatles, YOUTUBE
(Sep. 9, 2016), https://www.youtube.com/watch?v=LSHZ_b5W7o.
0
Id.

2
Even though those outcomes might create an impression of creative work at first sight, this
Article will show that there are still significant differences between artificial and natural
intelligences, notably in that Al is not capable of creative thinking. Hence, AI's work does not
qualify for copyright protection.

I. Anthropocentric Copyright

Despite all the variations in the world's multitude of copyright regimes, a common thread
emerges: copyright is based on an anthropocentric perspective. Authors are considered to be
human. In the United States, this has always been the courts' understanding of the U.S.
Constitution's wording: "By securing for limited Times to Authors . . the exclusive Right to
their respective Writings."" Recently, in a widely noted decision, the U.S. Court of Appeals for
the Ninth Circuit dismissed a copyright infringement claim brought by an institution on behalf of
a monkey that sought to assert rights in a self-clicked photograph, thereby affirming authorship
as a right exclusive to humans.12 Furthermore, the U.S. Copyright Office requires human
authorship 13 and "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." 14 Hence, the proposition that the U.S. Copyright Act does not require human
authorship, as some proponents of copyright protection for Al products contend, 5 appears to be a

" U.S. CONST. art I, § 8, cl. 8; see, e.g., Burrow-Giles Lithographic Co. v. Sarony, 111 U.S. 53, 61 (1884) ("[A]nd
Lord Justice BOWEN says that photography is to be treated for the purposes of the act as an art, and the author is
the man who really represents, creates, or gives effect to the idea, fancy, or imagination."); Urantia Found. v.
Maaherra, 114 F.3d 955, 958 (9th Cir. 1997) (" [S]ome element of human creativity must have occurred in order for
the Book to be copyrightable."); Kelley v. Chi. Park Dist., 635 F.3d 290, 304 (7th Cir. 2011) ("Authors of
copyrightable works must be human; works owing their form to the forces of nature cannot be copyrighted.").
12 Naruto v. Slater, 888 F.3d 418, 420 (9th Cir. 2018) ("[W]e conclude that this monkey-and all animals, since they

are not human-lacks statutory standing under the Copyright Act."); id. at 426 ("17 U.S.C. §§ 101, 201, 203,
304.... § 203(a)(2)(A). The terms 'children,' 'grandchildren,' 'legitimate,' 'widow,' and 'widower' [in those
provisions] all imply humanity and necessarily exclude animals .... "). Contra Victor M. Palace, Note, What If
Artificial Intelligence Wrote This? Artificial Intelligence and CopyrightLaw, 71 FLA. L. REv. 217, 226 (2019)
(asserting that this judgment has no impact on Al as this technology was not subject to this case).
13 U.S. COPYRIGHT OFF., COMPENDIUM OF U.S. COPYRIGHT OFFICE PRACTICES § 306 (3d ed. 2017),
https://www.copyright.gov/comp3/chap300/ch300-copyrightable-authorship.pdf ("The U.S. Copyright Office will
register an original work of authorship, provided that the work was created by a human being.").
" Id. § 313.2. The registration with the Copyright Office is governed by 17 U.S.C. §§ 408-12. In general,
copyright protection does not presuppose registration of a work. No civil action for infringement of a U.S. work,
however, will succeed without application for registration. 17 U.S.C. § 411(a).
15 See Annemarie Bridy, Coding Creativity: Copyrightand the Artificially IntelligentAuthor, 2012 STAN. TECH. L.

REv. 5 ¶ 49 (2012) ("Because copyright law does not expressly require human authorship, artificially intelligent
computer programs that autonomously generate art need not be relegated for copyright purposes to scare-quoted
authorship; their works can be regarded as proper 'works of authorship' under § 102 of the Copyright Act by virtue
of their nexus to human creativity."); Annemarie Bridy, The Evolution ofAuthorship: Work Made by Code, 39
COLUM. J.L. & ARTS 395, 399 (2016) ("The Copyright Act doesn't say anywhere that an author has to [sic]
human .... "); see also Margot E. Kaminski, Authorship, Disrupted:AI Authors in Copyright and FirstAmendment
Law, 51 UC DAVIs L. REv. 589, 603 (2017) ("[T]he U.S. copyright system has already moved far enough away
from romantic authorship for algorithmic authorship .... "); Yanisky-Ravid, supra note 1, at 718-19 ("It is not clear
whether the U.S. Copyright Act itself explicitly requires the author of a creative work to be a human.").
minority view, not founded in case law or legal practice. 16 It is clear that Congress did not intend
to grant intellectual property rights to machines." The situation is even more unambiguous in
Europe, where works are described as an "author's own intellectual creation." The European
Directive on computer programs stipulates that a "computer program shall be protected if it is
original in the sense that it is the author's own intellectual creation. No other criteria shall be
applied to determine its eligibility for protection." 18 The directive on databases as well as a
directive on certain copyright aspects contain the same provision. 19 Recital 16 of the latter
directive further clarifies that a "photographic 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." 20 Based on these provisions, the European Court of Justice
(ECJ) unified the protection requirements for all works, characterizing a work as an "author's
own intellectual creation"2 1 and explaining that "the author . .. can stamp the work created with
his 'personal touch,"' 22 thereby implying human authorship as a prerequisite. 2 3 This position
corresponds with the legal status quo in Switzerland2 4 as well as, for example, the prevailing

16 James Grimmelmann, There's No Such Thing As a Computer-Authored Work And It's a Good Thing, Too, 39
COLUM. J.L. & ARTS 403, 403 (2016) ("[Computer-authored works] don't exist. Copyright law doesn't recognize
computer programs as authors, and it shouldn't."); Tim W. Dornis, Artificial Creativity: Emergent Works and the
Void in CurrentCopyright Doctrine, 22 YALE J.L. & TECH. 1, 20 (2020) ("U.S. copyright ... is still based on the
concept of a natural person being the author or creator and, accordingly, the right-holder."); Daniel J. Gervais, The
Machine as Author, 105 IOWA L. REv. 2053, 2068 (2020) ("As applied by courts, originality requires human
authorship.").
" Pamela Samuelson, Allocating OwnershipRights in Computer-GeneratedWorks, 47 U. PITT. L. REV. 1185, 1199
(1986) (" [I]t is still fair to say that it was not within Congress' contemplation to grant intellectual property rights to
machines. In the long history of the copyright system, rights have been allocated only to humans.").
18 Directive 2009/24, of the European Parliament and of the Council of 23 April 2009 on the Legal Protection of

Computer Programs, art. 1 ¶ 3, 2009 O.J. (L 111) 16, 18 (EC).

19 Directive 96/9, of the European Parliament and of the Council of 11 March 1996 on the Legal Protection of
Databases, art. 3 ¶ 1, 1996 O.J. (L 77) 20, 25 (EC) ("In accordance with this Directive, databases which, by reason
of the selection or arrangement of their contents, constitute the author's own intellectual creation shall be protected
as such by copyright. No other criteria shall be applied to determine their eligibility for that protection."); 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) ("Photographs which are original in the
sense that they are the author's own intellectual creation shall be protected . . .. No other criteria shall be applied to
determine their eligibility for protection.").
20 Directive 2006/116, supra note 19, at 13
¶ 16.
21 Case C-5/08, Infopaq Int'l A/S v. Danske Dagblades Forening, 2009 E.C.R. I-6624, ¶¶ 35-38.
22 Case C-145/10, Painerv. Standard VerlagsGmbH, 2011 E.C.R. I-12594, ¶ 92; see also Case C-604/10, Football

Dataco Ltd. v. Yahoo! UK Ltd., ECLI:EU:C:2012:115, ¶ 38 (Mar. 1, 2012) ("[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 de Cock Buning, supra note 6, at 314; Jani Ihalainen, Computer Creativity:Artificial Intelligence and Copyright,
13 J. INTELL. PROP. L. & PRAC. 724, 727 (2018).
24 See URHEBERRECHTSGESETZ [URG] [COPYRIGHT ACT], Oct. 9, 1992, AS 1798 (1993), art. 6 (Switz.) ("The

natural person who has created a work shall be deemed the author."); see also Tim Rohner, Der Schutz von KI-
Schopfungen im schweizerischen Urheberrecht, 11 INTELL. PROP. J. 33, 53-69 (2019) (Ger.); Daniel Sch6nberger,
Deep Copyright: Up- and Downstream QuestionsRelated to Artificial Intelligence (Al) and Machine Learning
(ML), 10 INTELL. PROP. J. 35, 45 (2018).

4
view in Australia" and Japan. 2 6 Moreover, the Berne Convention, the oldest international
agreement governing copyright, implies that human authorship is necessary for copyright
protection. 27 The World Intellectual Property Organization (WIPO), while discussing a Model
Copyright Law, considered including "computer-produced works" in that model law,2 8 but never
actually implemented it. Additionally, many jurisdictions deny patent protection to non-human
creators as well. Both the U.S. Patent and Trademark Office (USPTO) and the European Patent
Office (EPO) refuse to accept a machine as an inventor, the latter invoking an internationally
applicable standard of inventors being natural persons referring, inter alia, to the U.S. Court of
Appeals for the Federal Circuit as wells as the patent offices of China, Japan, Korea, and the
USA. 2 9

Under this approach, AI-created works remain unprotected because the outcome stems from such
an independent process that it cannot be attributed to any human being. As the algorithms
improve and refine themselves, the "Human Behind the Machine" is moving more and more into
the background as the Al software is operating more and more independently. 3 0 Whether a work

25 Cf Acohs Pty Ltd v Ucorp Pty Ltd [2010] FCA 577 (10 June 2010) 66 (Austl.) (holding that source codes
generated by software were not original works by authors of software and thus not protected by copyright laws);
Telstra Corp. v Phone DirectoriesCo Pty [2010] FCA 44 (8 February 2010) 102 (Austl.), aff'd, [2010] FCAFC 149
(15 December 2010) (denying copyright protection in a work of which "substantial parts ... do not have human
authors ... ,are automated to the extent that human involvement is minor ... ,or have authors who cannot be
ascertained .... "); see also Ihalainen, supra note 23, at 726 ("Australian legislation ... doubts the existence of
copyright in works created by Al, even with human input.").
26 See Takashi B. Yamamoto, Al Created Works and Copyright, 48 PATS. & LICENSING, no. 1, 2018, at 1, 3,
https://www.itlaw.j p/AIp.
20Created%20Works%20and%2OCopyright.pdf ("The critical requirement 'creativity' here is construed by the court
cases that the personality of an author is revealed in expression in any way. Accordingly, only human created works
may fall within the protectable works.").
27 See Jane C. Ginsburg, People Not Machines: Authorship and What It Means in the Berne Convention, 49 INT'L
REV. INTELL. PROP. & COMPETITION L. 131, 134-35 (2018) ("[B]ecause lack of human authorship would disqualify
such outputs from Berne subject matter under Art. 2, other Berne members incur no obligation to protect purely
computer-generated works even if their countries of origin choose to cover them by copyright."). For example,
Art. 3 of the Berne Convention for the Protection of Literary and Artistic Works of Sept. 9, 1886, last amended on
Sept. 28, 1979, deals with nationality and habitual residence, neither of which would apply to an Al author.
28 WORLD INTELL. PROP. ORG., COPYRIGHT: MONTHLY REVIEW OF THE WORLD INTELLECTUAL PROPERTY
ORGANIZATION (WIPO), at 258 (1990),
https://www.wipo.int/edocs/pubdocs/en/copyright/120/wipo-pub_ 120_1990_09.pdf.
29 See In re Application of Flashpoint IP Ltd., No. 16/524,350 (U.S.P.T.O. July 29, 2019),
https://www.uspto.gov/sites/default/files/documents/165243 50_22apr2020.pdf?utmcampaign=subscriptioncenter&
utm_content=&utm_medium=email&utm_name=&utm_source=govdelivery&utm_term=; see also European Pat.
Off., Groundsfor the EPODecision of 27 January2020 on EP 18 275 174 (Jan. 27, 2020),¶ 30,
https://register.epo.org/application?documentld=E4B630BI2076498&number=EP 18275174&lng=en&npl=false
(referring to Univ. of Utah v. Max-Planck-Gesellschaft, 734 F.3d 1315, 1323 (Fed. Cir. 2013)). For a further
discussion on the EPO decision, see Martin Stierle, Artificial Intelligence Designatedas Inventor -An Analysis of
the Recent EPO Case Law, 69 GRUR INT'L: J. EUROPEAN & INT'L IP L. 918 (2020).
30 Rohner, supra note 24, at 52; see also Ginsburg, supra note 27, at 133-34. ("[H]uman users do not contribute

sufficient 'intellectual creation' to meet minimum standards of authorship under the Berne Convention. Offline,
merely giving a command does not make one an 'author': Pope Julius II may have commissioned the painting of the
ceiling of the Sistine Chapel; from a Berne perspective (at the very least), the author of the frescos remains

5
created with the help of Al as a "computer-assisted work" is protectable depends, roughly
speaking, on whether Al was only employed as a tool for implementing human decisions or
acted independently.31

At the top of the list of countries taking the opposite approach is the United Kingdom, which
provides for the protection of "computer generated works," that is, works that are "generated by
computer in circumstances such that there is no human author of the work." 32 South Africa,3 3
New Zealand,3 4 Ireland," and India 6 also follow this standard. The first country to protect
computer-generated works, the United Kingdom, had already assumed, at the end of the 1980s,

Michelangelo."); Dornis, supra note 16, at 8 ("[T]he AI's actual functions, operation, and productive output are
determined entirely by its acquired capacities. With such an alteration of the process comes a loss of predictability,
and Al autonomy ensues."); Palace, supra note 12, at 236 ("Thus, [the user, programmer, and company] would be
rewarded despite not contributing to the intellectual conception of the work, contrary to the purpose of copyright
law."); Robert Yu, Comment, The MachineAuthor: What Level of CopyrightProtectionIs Appropriatefor Fully
Independent Computer-generatedWorks?, 165 U. PA. L. REv. 1245, 1258-59 (2017) (explaining that a machine-
generated work is neither a derivate work of the Al software since the work is not predicated on the previous work
nor the end-user's creation since the end-user does not determine composition or arrangement); Gervais, supra note
16, at 2070 ("The automated decision-making feature of deep learning machines ... adds unpredictability-but not
randomness-and in doing so it breaks the causal link between humans ... and the output.").
31 Kalin Hristov, Artificial Intelligence and the CopyrightDilemma, 57 IDEA 431, 435 (2017) ("In this category [of

Al programs with the direct guidance, assistance or input of human beings], Al is used as a tool to achieve a
determined or predicted goal or outcome. An example may be the creation of a painting by an artist who has selected
the colors, tool type . .. and has to some extent input his requirements into the Al algorithm used to create the
work."); Dornis, supra note 16, at 7, 22 ("[A] distinction is required as to whether it merely functions as a tool or
instrument of the human actor's creativity, or whether the human actor has given up control over both the production
process and its outcome."); see generally Jane Ginsburg & Luke Ali Budiardjo, Authors and Machines, 34
BERKELEY TECH. L.J. 343 (2019).
32 Copyright, Designs and Patents Act 1988, c. 48, § 178 (UK). According to the High Court of England and Wales
(High Court of Justice), composite screen frames, generated by a computer program, of a coin operated video game
are computer-generated works because the software built up composite images by overlaying the digital image of a
pool table with images of the balls and cue. See Nova Productions v. Mazooma Games, [2006] EWHC 24 (Ch)
(UK). With respect to authorship, § 9(3) reads: "In the case of a literary, dramatic, musical or artistic work which is
computer-generated, the author shall be taken to be the person by whom the arrangements necessary for the creation
of the work are undertaken." Id. And the scope of protection is somewhat limited by excluding computer-generated
works from moral rights and shortening the duration of copyright. See Copyright, Designs, and Patents Act 1988, c.
48 §§ 12(7), 79(2), 81(2).
33 Copyright Act 98 of 1978 § 1(h) (S. Afr.) ("In this Act, unless the context otherwise indicates, 'author', in relation
to a literary, dramatic, musical or artistic work or computer program which is computer-generated, means the person
by whom the arrangements necessary for the creation of the work were undertaken [[A]dded by s. 1 (d) of Act No.
125 of 1992].").
34 Copyright Act 1994, ss 2, 5(2)(a) (N.Z.) (stating that "the person who creates a work shall be taken to be in the
case of a literary, dramatic, musical, or artistic work that is computer-generated, the person by whom the
arrangements necessary for the creation of the work are undertaken").
35 Copyright and Related Rights Act 2000 (Act. No. 28/2000) (Ir.), Part I, § 2, § 21(f),
http://www.irishstatutebook.ie/eli/2000/act/28/enacted/en/html (stating that an "'author' means the person who
creates a work and includes in the case of a work which is computer-generated, the person by whom the
arrangements necessary for the creation of the work are undertaken").
36 Indian Copyright Act, 1957, § 2(d)(vi) ("'[A]uthor' means, in relation to any literary, dramatic, musical or artistic

work which is computer-generated, the person who causes the work to be created.").

6
that "the far-sighted incorporation of computer-generated works in [its] copyright system will
allow investment in artificial intelligence systems, in the future, to be made with confidence." 37

On December 24, 2019, a Chinese court conferred copyright protection on an AI-written text. 38
However, it did so without relying on the fact of pure machine creation. Instead, it based its
reasoning on the fact that the prerequisite of "intellectual creation" under Chinese copyright law
was met by means of human involvement to a certain-albeit small-extent through
determining, among other things, selection and arrangement. 39 Thus, at least in this court's view,
the case did not concern a work entirely created by Al. 40

Contemplating both national jurisdictions as well as international law, it becomes clear that
copyright law is designed to be anthropocentric, admittedly with the exception of computer-
generated works in a few countries. Proponents of exclusive rights for Al products either urge
the United States to follow their example and join the United Kingdom, 4 1 or propose to apply the
work-made-for-hire doctrine, 42 considering the Al technology to be its programmer's or user's
employee. 43 However, postulating such a flexible definition of "employer" and "employee" in
the U.S. Copyright Act's definition of "work made for hire" 4 4 blatantly conflicts with the U.S.
Supreme Court's holding that the term "employee" must be interpreted in accordance with
agency law. 45 It should be generally and internationally recognized that machines are owned, not
employed. 46

37 HL Deb (12 Nov. 1987) (489) col. 1521 (UK), https://api.parliament.uk/historic-

hansard/lords/1987/nov/12/copyright-designs-and-patents-bill-hl#column_1476 (quoting the Earl of Stockton
Maurice Harold Macmillan).
38 Shenzhen Tencent Comput. Sys. Co. v. Shanghai Yingxun Tech. Co., (People's Ct. of Nanshan (Dist. Of

Shenzhen) Dec. 24, 2019) (China), translatedin 51 INT'L REV. INTELL. PROP. & COMPETITION L. 652 (2020).
39
Id.at 657.
40 Zhou Bo, Artificial Intelligence and CopyrightProtection-JudicialPractice in Chinese Courts, WORLD INTELL.
PROP. ORG., https://www.wipo.int/export/sites/www/about-
ip/en/artificial_intelligence/conversation_ipai/pdf/ms_china_1_en.pdf (last visited Nov. 29, 2020).
41 E.g., Robert Denicola, Ex Machina: CopyrightProtectionfor ComputergeneratedWorks, 69 RUTGERS U. L. REv.

251, 287 ("The United States, either by judicial decision or statutory amendment, should join them."); Kasap, supra
note 1, at 377 (" [T]he CDPA is a reasonable pathway that the United States legislature should follow .....
42 17 U.S.C. § 201(b).
43 See Bridy, Coding Creativity, supra note 15, at ¶ 66; Hristov, supra note 31, at 442-43; see also Yanisky-Ravid,
supra note 1, at 712 ("The autonomous Al system, just like WMFH-employed creators, is the creative author of a
work. When an Al system acts autonomously, it can be compared to an independent contractor and thus be shielded
under WMFH doctrine.").
44 17 U.S.C. § 101. See Hristov, supra note 31, at 446 ("'[E]mployer' may be considered as someone who employs
the services of another entity in order to achieve a goal or complete a task.").
41 Cmty. for Creative Non-Violence v. Reid, 490 U.S. 730, 739-41 (1989).
46 European Pat. Off., supra note 29,
¶ 32.

7
II. Creativity as Copyright's Basis

Regardless of the legal status quo in the United States, it is to be examined whether the human
creation requirement should be abolished, or the "work made for hire" doctrine should be more
generously interpreted for policy reasons.

Copyright law is not only anthropocentric but also "intended to motivate the creative activity of
authors and inventors." 47 Its "ultimate aim is . . . to stimulate artistic creativity for the general
public good." 48 According to the U.S. Supreme Court, the originality requirement in the
Copyright Act49 necessitates "that the author make the selection or arrangement independently
(i.e., without copying that selection or arrangement from another work), and that it display some
minimal level of creativity." 0 Similarly, the ECJ requires that an author "express his creative
abilities in the production of the work by making free and creative choices."" The Chinese court
case discussed in Part I also took into account a "creative process" and based its argument on
adjudging a creative act."

This illustrates that, despite all the remarkable differences between the U.S. copyright framework
and other copyright regimes, especially in civil law countries, the creativity-based approach
constitutes a global principle.5 3 In fact, the author of a work is traditionally more central in civil

4 Sony Corp. of Am. v. Universal City Studios, Inc., 464 U.S. 417, 429 (1984).
48 Twentieth Century Music Corp. v. Aiken, 422 U.S. 151, 156 (1975).
49 17 U.S.C. § 102.
50 Feist Publ'ns, Inc. v. Rural Tele. Serv. Co., 499 U.S. 340, 358 (1991); id. at 363 ("[C]opyright protects
only those
constituent elements of a work that possess more than a de minimis quantum of creativity."); see also Goldstein v.
California, 412 U.S. 546, 561 (1973) ("[T]he word 'writing' may be interpreted to include any physical rendering of
the fruits of creative intellectual or aesthetic labor.").
51 Case C-145/10, Painerv. Standard VerlagsGmbH, 2011 E.C.R. I-12594, ¶ 89; see also Case C-604/10, Football

Dataco Ltd. v. Yahoo! UK Ltd., ECLI:EU:C:2012:115, ¶ 38 (Mar. 1, 2012).

52 Shenzhen Tencent Comput. Sys. Co. v. Shanghai Yingxun Tech. Co., (People's Ct. of Nanshan (Dist. Of
Shenzhen) Dec. 24, 2019) (China), translatedin 51 INT'L REV. INTELL. PROP. & COMPETITION L. 652, 657 (2020)
("[W]hen determining whether an act is creative, it should be considered whether the act is an intellectual activity
and whether there is a direct connection between the act and the specific expression of the works.").
53 Dornis, supra note 16, at 17 ("Copyright protection in virtually all jurisdictions depends on the quintessential
element of human creativity. Both civil-law and common-law copyright have an anthropocentric foundation.");
Sch6nberger, supra note 24, at 37 ("[I]n the existing international copyright landscape creativity is the touchstone
for any literate or artistic work seeking protection. Since the Supreme Court decision in Feist, US contemporary
copyright law has been understood as 'creativity law'. And also copyright regimes in the droit d'auteurtradition put
intellectual creativity at the heart of the protection provided."). The United Kingdom, on the other hand, which is at
the forefront of protecting computer-generated works, dissociates itself from a creativity-based approach. See
Andreas Rahmatian, Originalityin UK Copyright Law: The Old "Skill and Labour" Doctrine Under Pressure, 44
INT'L REV. INTELL. PROP. & COMPETITION L. 4, 13 (2013) ("It is not (artistic) creativity, but the potential (not
actual) economic value of the author's investment, skill and labour deployed in the making of a property (the
copyright work) which copyright protects."); PAUL TORREMANS, HOLYOAK AND TORREMANS INTELLECTUAL
PROPERTY LAw 178-79 (9th ed. 2019) (" [T] here must have been a minimum investment by the author of skill,
judgement, and labour .... On the one hand, that skill and labour must not be so trivial that it could be
characterized as a purely mechanical exercise; on the other, creativity, as such, is not required either.").

8
law than in common law, which tends to emphasize social and economic welfare.5 4 Since civil
law countries emphasize moral rights, authorship of machines is even more difficult for civil law
regimes to justify than for common law regimes. 5 5 Notwithstanding different emphases,
however, both systems are based on the notion that works are public goods; that is, they may be
exploited by an unlimited number of users, but do not allow for remuneration because others
cannot be excluded. 56 This discourages creation of works, resulting in a market failure. To
offset this, copyright systems give authors an incentive to produce and publish their works, thus
making them open to the public.57

Worldwide, a common thread can be discerned: creativity is copyright's basis. Many scholars
assume that Al is creative in this sense. 58 However, the devil is in the details.

5 Ginsburg, supra note 27, at 134 ("Copyright, however, reposes on two pillars (whose respective widths vary in
common law and civilian systems): one (generally attributed to civil law states), the natural rights of the author, a
rationale that roots exclusive rights in personal creativity, and that largely underpins the Berne Convention; the other
(most frequently associated with common law countries), incentives to create, to invest in creativity, and to
disseminate works for the general benefit of society."); see also de Cock Buning, supra note 6, at 319-20 ("The
Anglo-American copyright law traditionally has a more pragmatic approach with regard to authorship, since it is
vested in the US Constitution as an incentive for creation and innovation.. . . It puts less emphasis on the protection
of the creator/author and more on the furthering of production of works that have value for society. As a result, in
the United States - like in the United Kingdom - the resistance against protection of (partial) machine creation is
traditionally less fierce.... Given the standard of harmonization of European copyright law, this is and will be the
leading approach towards copyright in all Member States of the European Union, including the UK, leaving little or
no room for the protection of the output of [Autonomous Intelligent Systems] as Creative Agents."); Matthias
Leistner & Gerd Hansen, Die Begrundung des Urheberrechtsim digitalen Zeitalter - Versuch einer
Zusammenfihrung von individualistischenund utilitaristischenRechtfertigungsbemuhungen, GEWERBLICHER
RECHTSSCHUTZ UND URHEBERRECHT 479 (2008) (Ger.) (arguing that, also in the digital age, only a synthesis of
individualistic and utilitarian approaches may justify a copyright in continental European design); Rohner, supra
note 24, at 72 (demonstrating that, historically, German-speaking scholars stressed the personal and idealistic
component of copyright); Yanisky-Ravid, supra note 1, at 699 ("Today, U.S. intellectual property law is based
primarily on the law and economics utilitarianism approach and, in part, John Locke's theory of labor. By contrast,
the civil law approach to copyright protection justifies property rights by the importance of the creators' personality
in the works (personality approach), as well as by the ownership of the fruits stemming from the person's body and
soul (Locke's approach or labor approach.").
5 Bridy, supra note 15, at 400-01. Moral rights concern the intellectual and personal relationship of the author to
their work, i.e. rather spiritual than economic interests. Martin Miernicki & Irene Ng (Huang Ying), Artificial
Intelligence andMoral Rights, Al & Soc'Y (Aug. 3, 2020), https://doi.org/10.1007/s00146-020-01027-6 ("Moral
rights acknowledge that authors have personal interests in their creations and the corresponding use that is made of
them. These interests are conceptually different from the economic or commercial interests protected by the author's
economic rights which are typically understood to enable the author to derive financial gain from her creation
.... ").
56 TORREMANS, supra note 53, at 18; Rohner, supra note 24, at 74.
" See TORREMANS, supra note 53, at 17-18;Yanisky-Ravid, supra note 1, at 700; Domis, supra note 16, at 33-36;
Colin R. Davies, An Evolutionary Step in Intellectual Property Rights- Artificial Intelligence and Intellectual
Property, 27 COMPUT. L. & SEC. REV. 601, 605 (2011); see also Mazer v. Stein, 347 U.S. 201, 219 (1954) ("The
economic philosophy ... is the conviction that encouragement of individual effort by personal gain is the best way
to advance public welfare through the talents of authors and inventors .... ").
58 E.g., Hristov, supra note 31, at 434 ("[C]reativity machines and other forms of Al will likely take center stage in
the creative process, becoming the main drivers of creativity and innovation."); Denicola, supra note 41, at 272
("'[T]he requisite level of creativity is extremely low; even a slight amount will suffice.' Many computer-generated

9
 A. Outlining Creativity

Everyone has an idea of what creativity is. Nevertheless, it is difficult to outline it, as there is no
universal definition. 59 Therefore, this Article can only demonstrate certain aspects of creativity
which academia discusses. Most importantly, in many definitional approaches, creative ideas are
characterized both as new or novel as well as valuable 60 or adaptive. 6 1 Philosophers and
psychologists essentially agree on this. 62

Generally, creativity manifests itself in three ways: (1) as an intellectual process that produces
adaptable and original ideas, (2) as a kind of person who carries creativity to the outside world,
and (3) as the concrete work that emerges from the creative process. 63 Typically, character traits,
biographical experiences, and emotions, among other things, serve as sources of creative
thinking. 64 Hence, some require skill, appreciation, and imagination as a benchmark for
creativity.61

B. AI's Mirage

One could be forgiven for inferring, based on the rapidly increasing sophistication of AIs, that Al
is capable of creativity. That is because algorithms have amazing capabilities that-in biological
organisms-require complex cognitive processes that have only been observed in humans. For
example, algorithms can see items, understand languages, and draw conclusions. 66 In 2016, the

works easily meet that standard." (quoting Feist Publ'ns, Inc. v. Rural Tel. Serv. Co., 499 U.S. 340, 345 (1991));
Yanisky-Ravid, supra note 1, at 679 ("Al systems are capable of more than just copying other works from
accessible sources. They operate as creative devices capable of creating entirely new and original works."). But see
Yanisky-Ravid, supra note 1, at 724 ("It may be that the process by which an Al system creates an original work is
not 'creativity,' which, as a term, has not been thoroughly explained by the [U.S. Supreme] Court.... But it may
just as well be the case that the creativity standard the Court articulated in Feist requires that innate, hard-to-define
aesthetic sensibility that is, particular to living creatures."). Similarly, Boden deems the question of AI's creativity
"currently unanswerable." Margaret A. Boden, ComputerModels of Creativity, 30(3) Al MAG. 23, 33 (2009); see
also Dornis, supra note 16, at 15-16 (though "the results of such artificial creativity qualify as entertaining,
informative, and inspiring-all traits that were once considered as exclusively belonging to humans," Dornis
grounds his reasoning mainly on the fact that AI's skills are improving, so "[e]ven though Al may never perfectly
match all human capacities, the gap between artificial and natural intelligence will someday shrink beyond
recognition").
59 More than fifty definitions of creativity can be derived from different sources. DAVID LEVY, ROBOTS UNLIMITED:
LIFE IN A VIRTUAL AGE 149 (2006).
60 Boden, supra note 58, at 24.
61 See de Cock Buning, supra note 6, at 315.
62
See id. at 315-16.
63 Id. at 315. Similarly, the famous American educational scientist James Melvin Rhodes described creativity using

the "4Ps," Person, Process, Press, and Product. Aleksandra Gruszka & Min Tang, The 4P's CreativityModel and Its
Application in Different Fields, in HANDBOOK OF THE MANAGEMENT OF CREATIVITY AND INNOVATION: THEORY
AND PRACTICE 51 (Min Tang & Christian H. Werner eds., 2017).
64 Sarah Legner, Erzeugnisse KinstlicherIntelligenz im Urheberrecht,ZEITSCHRIFT FUR URHEBER- UND

MEDIENRECHT [ZUM] 807, 809-10 (2019) (Ger.).

65 Simon Colton, Creativity Versus the Perception of Creativity in ComputationalSystems, in AAAI SPRING
SYMPOSIUM: CREATIVE INTELLIGENT SYSTEMS 14 (2018), https://www.aaai.org/Papers/Symposia/Spring/2008/SS-
08-03/SS08-03-003.pdf.
66 de Cock Buning, supra note 6, at
312.

10
Al program AlphaGo even beat the best human player 67 of the board game Go. Go is arguably
the most complex board game and is a game that cannot be played solely by computing possible
moves. 68 These achievements could be viewed as steps towards the comprehensive superiority
of machines.

However, these spectacular achievements conceal AI's shortcomings relative to other areas of
human cognition. For instance, algorithms cannot plan and take initiative well. 69 For the time
being, machines depend on human instructions and leadership. 70 But most notably, they lack
internal understanding and awareness of what they are doing. 7 1 Machines do not reflect the
zeitgeist, do not process social and societal impressions, and do not get inspired on subconscious
levels. 7 2 Yet, according to the U.S. Supreme Court, this is a crucial factor for copyright
protection. 73 The mere fact that Al technology has the ability to surprise us and even those who
programmed and trained it does not necessarily amount to creativity and deserve authorship. 74

Regarding the project "You Can't Know My Mind," 75 the software's "mood" was merely a mock
mood influenced by data and, thus, just created the illusion of creativity. It is similar to the
"next" Rembrandt painting.76 Executive Bas Korsten, who had the idea for the project and
supervised it, admitted that the project was not intended to create a new Rembrandt.77 Therefore,
the result cannot be considered original creativity, but rather the absorption of Rembrandt's
creative output and, thus, a mere summary of this creativity into a supposedly new work. There
can be no doubt that the concept was innovative and exciting. However, it created a "typical

67 This player was Lee Sedol, the winner of eighteen world titles at that time. AlphaGo, DEEPMIND,
https://deepmind.com/research/case-studies/alphago-the-story-so-far (last visited Nov. 29, 2020).
68 See Paul Mozur, Google 's AlphaGo Defeats Chinese Go Master
in Win for A.I, N.Y. TIMES (May 23, 2017),
https://www.nytimes.com/2017/05/23/business/google-deepmind-alphago-go-champion-defeat.html; AlphaGo,
supra note 67; see also Lim, supra note 3, at 830.
69 de Cock Buning, supra note 6, at 313; see also Ginsburg & Budiardjo, supra note 31, at 394 ("[T]he idea of true
machine thought, guided by the sort of 'intrinsic motivation' that drives all human behavior, may still be far off.").
70 Ginsburg & Budiardjo, supra note 31, at 400 ("[I]t should suffice to note that today's machines, and those of
foreseeable tomorrows, are entirely subservient to the humans who delineate their instructions and tasks."). Al is,
hence, unlikely to overcome established principles and, thus, achieve significant advancements. See infra Part V.
71 Yanisky-Ravid, supra note 1, at 724 ("Even if a machine could create a unique rendering of a subject, it is very
unlikely that Al system would understand what that subject is. It thus lacks the type of internal comprehension that
is generally reflected in the works of a human artist when they try to represent something more than the words on
the page or the paint on the canvas.").
72 Kristin Oswald, KI als Kultur-Geschdftsfahrerder Zukunft, Kultur Management (Aug. 20, 2018),
https://www.kulturmanagement.net/Themen/Kuenstliche-Intelligenz-und-Kreativitaet-KI-als-Kultur-
Geschaeftsfuehrer-der-Zukunft,2348 (Ger.) (last visited Nov. 29, 2020) (interview with Prof. Dr. Paul Lukowicz,
Scientific Director and Head of the Research Department Embedded Intelligence at the German Research Center for
Artificial Intelligence).
73 Cf Feist Publ'ns v. Rural Tel. Serv. Co., 499 U.S. 340, 362 (1991) ("[A]n author who claims infringement must
prove 'the existence of . .. intellectual production, of thought, and conception. "' (quoting Burrow-Giles
Lithographic Co. v. Sarony, 111 U.S. 53, 59-60 (1884))).
74 Ginsburg & Budiardjo, supra note 31, at 398.
75 See supra note 6 and accompanying text.
76 See supra notes 7-8 and accompanying text.
77 Brown, supra note 7 ("Korsten stressed the project was not trying to create a new Rembrandt. 'We are creating
something new from his work. Only Rembrandt could create a Rembrandt."').

11
Rembrandt" rather than a "new Rembrandt." The same holds true for the song "Daddy's Car,"
which is based on Beatles music. 78 Under our current state of the art, Al always relies on
sufficient data, instructions and specifications, whereas the human creative process is not
confined. Think of another example: a tree that stretches its branches to form a handsome crown
to maximize its light absorption seems to act in a new and useful way; nonetheless, it lacks
intention, wishes, and views, so it cannot act creatively.7 9 Creativity requires, besides novelty
and usefulness, "a relevant purpose . . . some degree of understanding . .. a degree of
judgement . . and an evaluative ability directed to the task at hand." 80 There is much to suggest
that individual experiences and traits of the author influence their creative output. 81 Machines
cannot yet replicate that.

Therefore, natural and artificial intelligences should not (yet) be equated. 82 It follows that, due to
its lack of creativity, AI's work does not qualify for copyright protection. Some authors contend
that "the creativity the Al displays flows either from the algorithm used to design and train it, or
from the instructions provided by the users operating it," 83 but this blurs the line between
computer-generated and computer-assisted works. 84 While Al software is learning and, thus,
improving, the person creating it is moving so far into the background that the products cannot
be attributed to that person anymore. 85 Given the ever-increasing amount of computing power
available and the immense resources being directed to the development of artificial intelligence,
an artificial form of creativity might be achieved in the future.8 6 However, for the foreseeable
future, it is inappropriate to extend copyright protections to works created by algorithms.

78 For a link to that song, see supra note 9.

79
8
de Cock Buning, supra note 6, at 316.
0
Id.
81 Anne Lauber-R6nsberg, Autonome ,Schopfung" - Urheberschaftund Schutzfdhigkeit, GEWERBLICHER
RECHTSSCHUTZ UND URHEBERRECHT [GRUR] 244, 251-52 (2019) (Ger.).
82 See id.; Ginsburg & Budiardjo, supra note 31, at 349, 393-400 ("Because computers today, and for proximate
tomorrows, cannot themselves formulate creative plans or 'conceptions' to inform their execution of expressive
works, they lack the initiative that characterizes human authorship.... No machine is itself a source of creativity.");
Kasap, supra note 1, at 350 ("Al has not possessed all features of humans yet, and there is still a long way to acquire
those skills .... "); Gervais, supra note 16, at 2093 ("[M]achines cannot make creative choices ... "); Ron Miller,
Artificial Intelligence Is Not As Smart As You (or Elon Musk) Think, TECH CRUNCH (July 25, 2017),
https://techcrunch.com/2017/07/25/artificial-intelligence-is-not-as-smart-as-you-or-elon-musk-think/ (noting that
"[s]hould researchers ever become more successful at developing generalized Al, this could change, but for now
there are things that humans can do easily that are much more difficult to teach an algorithm, precisely because we
are not limited in our learning to a set of defined tasks"). ContraDavies, supra note 57, at 604 ("The one thing
intellectuals are reluctant to accept is the concept of a machine possessing the ability for creative thought."); Tim W.
Dornis, Der Schutz kunstlicher Kreativitdtim Immaterialguterrecht,GEWERBLICHER RECHTSSCHUTZ UND
URHEBERRECHT [GRUR] 1252 (1254-55) (2019) (Ger.) (asserting that that Al is already emancipating itself from
human mastery).
83 Lim, supra note 3, at 842.
84 On that distinction, see supra note 31 and accompanying
text.
85 See sources cited supra note 30.
86 Cf de Cock Buning, supra note 6, at 313 ("The current [Autonomous Intelligent Systems],
however sophisticated
and unpredictable their output may sometimes be, in themselves do not yet constitute fully autonomous, general-

12
III. Economic Aspects of Exclusive Protection

Beyond copyright protection, a sui generis right may be considered.87 Such rights do not exist in
the United States yet. European Union (EU) law, on the other hand, provides for sui generis
rights for particular subject matters, such as databases.8 8 International copyright usually refers to
these rights as "related rights." 8 9 Sui generis rights are similar to copyrights in that they are
exclusive, but they do not require copyrightable subject matter.

A. Incentives and Market Failure

The most prominent argument in favor of an intellectual property (IP) right in AI-created works
is that it would provide developers and entities with an incentive to invest in and employ Al
technology. 90 A machine itself does not need any incentive; however, the incentive for
disclosure shall be provided for the humans behind the technology. 9 1 Only benefits reaching at
least the level of investment will avoid market failure due to underproduction of creative
works.9 2 In general, it does seem desirable to protect investments.9 3 After all, given that refining
Al technologies presupposes sufficient funding, and sui generis rights under European law are
rooted in safeguarding investments, 94 affording such a right for machine-authored works seems
an obvious consequence.

purpose artificial intelligent Creative Agents since these systems still lack important capabilities such as planning
and taking initiatives. Yet, they do not stand alone, and they herald the coming of the 'singularity', the
technological lift-off point, in which a combination of scientific and technical breakthroughs will lead to an
explosion of self improvinartificial intelligence, representing the nearly vertical phase of exponential growth that
occurs whenthe rate is so extreme that technology seems to be expanding at infinite speed. Along those lines, the
development of systems and machines that will be capable of autonomous creation seems to be inevitable.").
87 Dornis, supra note 16, at 17 ("The assignment of genuine authorial copyrights may still be reserved for human-
made creations. But this does not exclude alternative means of protection."). The concept of sui generis rights does
not require human creation. Id. at 33, 44-59.
88 See Directive 96/9, supra note 19, at 25-27 art. 7-11.
89 E.g., TRIPS: Agreement on Trade-Related Aspects of Intellectual Property Rights pt. II § 1 , Apr. 15, 1994,
Marrakesh Agreement Establishing the World Trade Organization, Annex 1C, 1869 U.N.T.S. 299. They are called
droits voisins in France, Leistungsschutzrechte in Germany, diritti connessi in Italy, and naburigerechten in the
Netherlands. Domis, supra note 16, at 44.
90 See Hristov, supra note 31, at 438-39; Dornis, supra note 16, at 43; Dornis, supra note 82, at 1258-59; see also

Denicola, supra note 41, at 273 ("At least for now, the production of computer-generated works requires human
beings to develop, improve, distribute, and use the computer technology and to disseminate the resulting output. The
incentive of copyright protection may play a role, large or small, in all of these human activities.").
91 See Hristov, supra note 31, at 444; de Cock Buning, supra note 6, at 322; Denicola, supra note 41,
at 273; Dornis,
supra note 16, at 35; Kasap, supra note 1, at 362; Lim, supra note 3, at 840-42.
92 Dornis, supra note 16, at 36; Dornis, supra note 82, at 1258.
93 Sven Hetmank & Anne Lauber-R6nsberg, Kunstliche Intelligenz - Herausforderungenfur das
Immaterialgterrecht,GEWERBLICHER RECHTSSCHUTZ UND URHEBERRECHT [GRUR] 574, 579-80 (2018) (Ger.).
94 See, e.g., Directive 96/9, supra note 19, at 22 ¶ 39 ("[T]his Directive seeks to safeguard the position of makers of

databases against misappropriation of the results of the financial and professional investment made in obtaining and
[collecting] the contents by protecting the whole or substantial parts of a database against certain acts by a user or
competitor.").
13
On counter, however, is that there may already be enough incentives to create through existing
protections. 95 Namely, there is "copyright or patent protection of the software programs, patent
protection of the specialized machinery to produce works of fine and applied art, and copyright
or (in the EU) sui generis protection of the database the software consults."9 6 Furthermore, if the
main reason for deploying machines to create works is their speed, as it is, for instance, with
authoring online news and other immediately consumable media, no copyright protection is
needed to provide economic incentives.9 7 Moreover, as an overall phenomenon, investment in
Al is currently increasing very rapidly-from $282 million in 2011 to $9.334 billion in 2018 in
the United States, 9 8 even though many relevant markets, including the United States and China,
do not provide protection. 99 That results in an average increase of about 165 percent per year.
Globally, investments in Al increased from $4.5 billion in 2013 to $39.2 billion in 2017.100 Not
surprisingly, the number of patent applications worldwide in the area of Al has risen from 22,913
in 2008 to 78,085 in 2018.101 The global revenues from the Al software market is expected to
grow from $10.1 billion in 2018 to $126 billion in 2025.102

The Al industry and Al users may invest in and employ that technology out of miscellaneous
motivations, which may not always include copyright protection. 103 The continuing increase in
investments might be explained by lower development costs, for example, when designing
fashion or writing articles, which is why a smaller incentive through mere software protection
suffices. In any case, a sui generis right would only be economically efficient if the resulting
benefit outweighed the costs. However, since public domain products can be used optimally and
there are barely any apparent obstacles to investment, there is much to suggest that demands for
an exclusive right are based rather on lobbying than on economic necessity. 10 4 Regardless, the
burden of proof lies with proponents of protection; whoever pushes for abridging the freedom of

95 Rohner, supra note 24, at 77; Yanisky-Ravid, supra note 1, at 702; see also Palace, supra note 12, at 239
("[T]here is little reason to believe that immediate entrance into the public domain would lead to any significant loss
in incentives for programmers and artificial intelligence companies.").
96 Ginsburg, supra note 27, at 134; see also Samuelson, supra note 17, at 1225 ("The programmer can already be
rewarded for the commercial value of the program-which is what he created-through sales of the program or
license fees for use of the program. That is probably all the motivation he needs.").
97 Yu, supra note 30, at 1246-47, 1264-65 (referring to an example of the L.A. Times, which in 2014 published an
online report on an earthquake just a couple of minutes after it had happened; software called Quakebot received the
geological data and immediately generated the article).
98 Shanhong Liu, Artificial Intelligence (Al) FundingInvestment in the United States From 2011 to 2019, STATISTA

(June 8, 2020), https://www.statista.com/statistics/672712/ai-funding-united-states/.

99 Rohner, supra note 24, at 77-78, 82 (asserting an increase of global investments by 300 percent from 2016 to
2017). Neither does the EU provide protection. See supra notes 18-23 and accompanying text.
100 Thomas Alsop, GlobalArtificialIntelligence (Al) Investment and FinancingFrom 2013 to Qi 2018, STATISTA
(Mar. 2, 2020), https://www.statista.com/statistics/941137/ai-investment-and-funding-worldwide/.
101 Shanhong Liu, Number ofArtificial Intelligence (Al) PatentApplications Worldwidefrom 2008 to 2018,
STATISTA (July 7, 2019), https://www.statista.com/statistics/1007843/number-of-ai-patent-applications-worldwide/.
102 Shanhong Liu, Artificial Intelligence (Al) Software Market Revenue Worldwide From 2018 to 2025, STATISTA

(Mar. 13, 2020), https://www.statista.com/statistics/607716/worldwide-artificial-intelligence-market-revenues/.

103 Yu, supra note 30, at 1270 ("Regardless of whether or not the law grants copyright protection, programs will

continue to work away with each passing moment, processing, computing, and creating.").
104 Cf Rohner, supra note 24, at 70.

14
use through establishing an exclusive right must prove its economic justification. 105 Market
failure is not proven yet. So, even if Al became what we deem creative one day, this would not
automatically justify a call for copyright protection.

B. Skepticism Towards Monopolies

In this author's view, other economic considerations militate against an exclusive right for AI-
generated works as well, be it a copyright or sui generis right. This is because by using Al,
works can be produced extremely quickly and in much larger numbers. Imagine a company
claiming to have generated almost every possible text of ten to 400 words in several common
languages and charging royalties for using them, thus becoming a huge copyright troll. In fact,
an Austrian performance artist did come up with that mock assertion in 2014.106 It would be
unlikely for such a business model to prevail as courts and legislators would never allow such an
absurd situation. Nevertheless, this made-up claim should startle us and make us aware of the
power of monopolies. Intelligent machines could drive the number of IP rights to a level far
beyond our current imagination. Too many IP rights-that is, government-granted
monopolies-lead to a restriction of competition, while at the same time the access to that
technology is limited to relatively few actors. 107 Hence, too many monopolies rather endanger
innovation, rendering the monopolies to be concentrated among those who are already powerful.
The powerful would, thus, become even more powerful. Only a small circle in just a few
countries would benefit enormously. 10 8 Should this occur, there is a threat of commercialization,
concentration, and homogenization of information.

IV. Supersession of Human Creation

Despite all the remarkable achievements and advantages of Al, another danger is looming.
Products created by Al provide the public with the same benefits as human creations. 109 Some
AI-generated products may even be perceived as superior in the market. 110 According to a study
from Rutgers University, respondents not only were unable to distinguish between computer-

105 See Ginsburg & Budiardjo, supra note 31, at 448 ("We can conjure up a variety of scenarios supporting or
debunking the call for sui generis protection, but without empirical evidence, it would be imprudent (and premature)
to seek to design a regime to cover authorless outputs.").
106 See Schafer et al., A Fourth Law of Robotics? Copyright and the Law and Ethics ofMachine Co-Production,23
A.I & L. 217, 225-26 (2015). See also Rohner, supra note 24, at 78-79. The artist's name is Michael Marcovici.
Schafer et al., supra note 106, at 226.
107 Hetmank & Lauber-R6nsberg, supra note 93, at 580; see also Rohner, supra note 24, at 78-79.
108 Palace, supra note 12, at
237.
109 Denicola, supra note 41, at 271; see also Dornis, supra note 16, at 42; Dornis, supra note 82, at 1260 (despite
supporting the notion of IP protection for Al generated works, acknowledging the issue of a possible substitution).
110 Hetmank & Lauber-R6nsberg, supra note 93, at 581.

15
generated and human art, but also ranked AI-created paintings as superior with regard to, among
other things, visual structure and inspiration."'

It is true that human creations remain protected because, regardless of higher or lower quality,
artistic quality is irrelevant in copyright law. Regardless, human beings could be economically
superseded by Al since the latter is capable of much great output, both in terms of rate and
quantity; this is truer if their products are perceived as superior. 112 Moreover, vesting protection
in Al products might discourage human artists from creating.1 13 Further, AI-created artistic
works may be seen as superior only in the short term; in the long term, however, works produced
by algorithms are likely to be more homogeneous, which would reduce diversity and turn
creative works into commodities.1 14 Ultimately, algorithms are not creative." 5 They are always
largely based on initial human instructions and what already exists, and consequently,
fundamental innovations are not to be expected. Al is capable of composing a new song in the
style of the Beatles 1 6 but probably could not revolutionize the music scene with a completely
new and astounding creation. It seems utterly questionable whether Al products could "promote
the Progress of Science and useful arts" as conceived by the U.S. Constitution.1 1 7 Therefore,
leaving the advantage of legal protection to human creation may prevent the diversion of too
much investment away from man-made works, thereby promoting essential innovation and
preserving copyright's major purpose.

This brings us to the question of whether the market is capable of regulating this issue itself by
reducing demand for (inferior) products of Al. However, if they are cheaper, the demand is
unlikely to be low. Furthermore, it is doubtful whether consumers can make sufficiently rational
and well-informed decisions by keeping an eye on the long-term development of crowding-out.

V. Recognizability and Transparency

Without any IP protection, it is contended, there is a danger that people will sell an AI-created
work as their own, refraining from disclosure that the work was actually created by algorithms,
which leads to a lack of transparency.is It would be more difficult for users to recognize

"ii Sarah Cascone, AI-GeneratedArt Now Looks More Convincingly Human Than Work atArt Basel, Study Says,
ARTNET (July 11, 2017), https://news.artnet.com/art-world/rutgers-artificial-intelligence-art-1019066; see also
Dornis, supra note 16, at 14.
112 ContraKasap, supra note 1, at 363 ("AI-systems should not be seen as replacing human authors, since obviously

they can evolve together."); Yanisky-Ravid, supra note 1, at 703 ("[JI]t is likely that machine-produced works could
not serve as a perfect replacement for human-authored works.").
113 Sch6nberger, supra note 24, at 46-47.
114 Ulrich Loewenheim & Matthias Leistner, PersonlicheSchopfung, in URHEBERRECHT
§ 2 UrhG ¶ 42 (6th ed.
2020) (Ger.).
115 See supra Part II.B.
116 See supra notes 9-10 and accompanying
text.
"1 U.S. CONST. art I, § 8, cl. 8.
118 See Hristov, supra note 31, at 450; Dornis, supra note 16, at 39; Dornis, supra note 82, at 1259; Kasap, supra

note 1, at 363.

16
whether the work is in the public domain or whether they need a permit. 119 Although, a certain
degree of vagueness is simply inherent in copyright law, given the ambiguity of the conditions
for protection-namely, the originality requirement under U.S. copyright law-registration with
the U.S. Copyright Office alleviates the transparency issue.12 0 Admittedly, applicants might still
make false statements as to a work's origin in order to receive protection. The same problem
might occur in other countries in cease-and-desist letters or during trials. However, this would
probably constitute an offense in many countries, and with respect to the United States, it should
also be noted that the use of Al can be discovered during litigation. 121

In addition, it could be argued that denying protection avoids another kind of problem of
ambiguity, namely of stipulating and determining on a case-by-case basis which natural person
owns the exclusive rights in an Al generated work. 12 2 Under UK law, it is "the person by whom
the arrangements necessary for the creation of the work are undertaken." 123 That leaves some
room for interpretation and controversy. This Article does not have to address that issue.

Conclusion

It is an internationally accepted principle of intellectual property law that copyright protection

attaches only to the products of human creativity. Responding to calls to extend intellectual
property rights to AI-generated works, this Article has raised significant economic and social
concerns that militate against overturning that global policy and granting intellectual property
protection for works produced by Al, whether it be in copyright or through a novel sui generis
right, in a civil law or common law jurisdiction. Whoever intends to establish a monopoly
through an exclusive right has to prove its economic efficiency and necessity. The more
convincing arguments prompt that existing incentives suffice. Companies already have reason to
employ and invest in Al technology in order to create products in typical copyright related fields,
such as music, design, literature, and software. Moreover, numerous new monopolies would
occur exceedingly fast, which might rather hamper than foster innovation and make the powerful
even more powerful. Finally, we should ask ourselves whether Al could overtake us and
supersede human creation, thus diminishing essential innovation.

Regardless, beyond those economic and social aspects, this Article contends that copyright
protection is quite the wrong approach for rewarding application of and investment in Al
technology. There can be no doubt that machines are superior in certain ways. However, in their

119 See, for example, Yu, supra note 30, at 1266, where even opponents of protection conceded that refraining from
disclosure will increase the difficulty of enforcing false copyrights.
120 See supra text accompanying notes
13-14.
121 Palace, supra note 12, at 237. Nonetheless, this problem of evidence is more substantial in jurisdictions without
discovery procedures. A conceivable alternative there might be labelling obligations under (European) unfair
competition law.
122 Right from the advent of the discussion on computer-generated works, that issue was on the table. Cf
Samuelson, supra note 17.
123 Copyright, Designs and Patents Act 1988, c. 48, @ 9(3)
(UK).

17
process of creating, they are still very distinguishable from humans. Algorithms lack crucial
skills, such as purpose, understanding, awareness, intuition, inspiration, and reflection. There
might be artificial intelligence, but there is no artificial creativity.

18