|
|
Home
| Databases
| WorldLII
| Search
| Feedback
University of New South Wales Law Journal Student Series |
THE IMPACT OF ARTIFICIAL INTELLIGENCE ON THE CULTURE INDUSTRIES AND COPYRIGHT LAW
JOLYN DE ROZA*
“The rise of the machines is here, but they do not come as conquerors, they come as creators”[1]
With rapid advancements in technology, artificial intelligence systems and technologies (collectively, “AI”) now encompass both automated learning and automated reasoning[2]. Automated learning refers to the process by which AI acquires information and applies predetermined rules for using that information; while automated reasoning refers to the process by which AI uses these rules to reach approximate or definite conclusions.[3] By applying such automated learning and automated reasoning, AI has been able to generate works independently or in a directed manner[4]. This essay is interested in the works which are generated independently by the AI, where the AI acts autonomously to create output that is unpredictable and undirected by human intervention[5]. For the purposes of this essay, such works are referred to as “AI generated works”.
The creation of AI generated works generally involves the following stages: (i) the programmers establish codes and software according to which the AI functions and feed the AI with data; (ii) users provide the parameters or instructions to direct the creation of the work; and (iii) the AI undertakes independent creative decisions to be the real originator of the final works[6].
AI generated works can be unique, complex and of artistic value[7]. One notable example is the Rembrandt Project, where computer programmers and art experts fed the AI with large volumes of data and established certain parameters around the portraits to be created. The AI was then able to create a painting in Rembrandt’s style[8]. Another highly publicised case is the AI generated artwork called “Portrait of Edmond Belamy” which was sold for US$432,500 in 2018.[9]
AI generated works go beyond art pieces. On the literary front – Future University Hakodate created an AI that wrote a novel, The Day a Computer Writes a Novel. This novel passed the first round of judging for the Hoshi Shinichi Award[10]. Although the University’s researchers selected the words, sentences and set parameters for the novel, the ultimate work submitted for the award was created by the AI.
AI generated works even include music. Flow Machines is an AI that is able to create music independently without human input[11]. The AI learns music styles from a huge database. A human user then selects a music style (for e.g. “American songwriters”), and the AI creates music for the user in the selected style. The Creativity Machine (an AI designed by Imagination Engines Inc) has created a full album, Song of the Neurons (available on iTunes) that is entirely composed and played by AI. No humans were involved in the musical output, and each note that followed “happened rather spontaneously.”[12]
These examples reflect the growing capacity of AI to generate creative works and monetary profits in the culture industries. This also raises pertinent issues for copyright law, in particular: (i) whether copyright should be attributed to AI generated works; and (ii) if copyright should be attributed to AI generated works, then in whom should the copyright vest?
Against this backdrop, Part II of this essay considers the current Australian copyright jurisprudence on AI generated works, and draws comparisons with the approaches adopted in the UK and the EU. Part II concludes that the current Australian approach requiring human authors to engage in independent intellectual effort is unlikely to protect AI generated works.
Part III of this essay then examines the policy arguments in favour of and against attributing copyright to AI generated works. Part III contends that the current Australian approach is not congruent with how works are created in this digital age, and argues in favour of granting copyright protection for AI generated works.
Part IV then suggests some legislative reforms for the protection of AI generated works. These suggested reforms are not intended to be exhaustive, but merely to generate discussion on the reforms which copyright laws could implement to address AI generated works.
Part V then concludes this essay.
Before moving to the next part of this essay, some points should be noted:
• There is a distinction between AI assisted works and AI generated works. AI assisted works generally refer to works where AI is merely a tool or a medium for the human author, like a brush in the hands of the painter to create the work.[13] As AI assisted works are unlikely to pose problems from a copyright perspective[14], this essay is only concerned with AI generated works.
• Apart from the creative works generated by AI (as illustrated above), there are other types of AI generated works, such as genetic programming and self-driving cars[15]. However, the focus of this essay is on the creative works relating to the culture industries (such as art, music and literature) that are generated by the AI.
• Since AI learns from the provided data, images and sounds to make its own predictions or decisions[16], a related copyright issue is whether the AI has infringed any copyright in the course of generating its work. However, such issues are not within the scope of this essay, as the focal point of this essay is on whether AI generated works should be protected by copyright.
The Copyright Act 1968 (Cth) (the “CA”) provides that “the author of a literary, dramatic, musical or artistic work is the owner of any copyright subsisting in the work”[17]. Although the CA does not provide any express definition for “author”, s 35(5) says that “an author” of a work can only be “a person” who has created the artwork. The CA s 208(1) also provides that “the author of a photograph shall ... be read as a reference to the person who, at the time when the photograph was taken, was the owner of the material on which the photograph was taken”. These provisions throughout the CA suggest that only humans can be identified as authors and owners of works.
Recent Australian cases have adopted a more restrictive approach to originality and authorship. In the earlier case of Desktop Marketing Systems v Telstra Corp[18] , the court had to consider whether telephone directories, which had been created through substantial use of computers, were sufficiently original to enjoy copyright protection. The Court found that the “industrious collection” directed to the creation of the directories was sufficient for a finding of originality, and held that it would be illogical to find that “the laborious and computer-aided compilation activity engaged in by Telstra” was insufficient to attract copyright protection.
However, this position was later reversed in IceTV Pty Limited v Nine Network Australia Pty Limited[19], where the Court considered the issue of copyright subsistence in a TV schedule. On the facts, the Court decided that there was minimal skill and labour in expressing the time and title for the creation of a TV schedule[20]. This decision had two significant implications. First, with respect to establishing originality for copyright subsistence, the Court held that only material originating from an author who has expended ‘independent intellectual effort’ will fulfil the requirement of originality[21]. Second, the Court found that although there was extensive human effort directed to making the programming decisions for the TV schedule, such work is only antecedent to the relevant act of authorship. The Court emphasised the role of human authorship in fixing the work in its material form. The lack of evidence as to how Nine’s database operated to select, arrange and present the format of the TV schedule meant that the author was unknown[22].
Based on the above, the current Australian approach may be regarded as adopting a higher threshold of originality and requiring a human to be seen as directing the material form of the work. In the absence of anything in the CA which makes provision for the authorship of computer-generated work, it follows from these cases that AI generated work will neither have the human authorship nor the element of intellectual creation, required to be protected by copyright.[23]
Indeed, more recent cases affirm that AI generated works are unlikely to be copyright protected.
Telstra Corporation Limited v Phone Directories Co Pty Ltd[24] concerned telephone directories that were created by an automated system which used a set of predetermined rules to conduct automatic checks of data and compile data into the directories. The Court found that copyright did not subsist in the telephone directories, because these directories had no human authors and were created by automated processes from databases. To the extent that individuals collected data and entered the data into the databases, such effort was not relevant to assessing authorship and whether sufficient intellectual effort had been undertaken. The court emphasised that the relevant activity to assessing these factors was the reduction of the telephone directories into their material form[25]. However, since the process of fixing the telephone directories into their material form was almost entirely automated, there was no identifiable author[26]. Lastly, the Court considered that software programmers of the automated system were not the authors of the telephone directories[27].
Finally, in the case of Acohs Pty Ltd v Ucorp Pty Ltd [28] , the Court had to decide whether copyright could be attributed to a HTML source code (for information sheets) which was generated by a computer programme. The court held that there was no copyright protection for the HTML source code, because it was not authored by a natural person[29]. In turning to consider whether the author of the computer programme could be the author of the HTML source code, the Court concluded that it would be artificial to regard the programmers in that way.[30]
Commentators have observed that AI generated works are unlikely to be deemed authored works under the Phone Directories formula, as the user is not “directing or fashioning the material form of the work”, and is comparable to a “human ‘on’ switch for random acts of creation”[31].
Most copyright legislation across the EU do not directly address AI generated works.[32] Hence, the issue of whether AI generated works can be protected by copyright will depend on the basic principles of awarding copyright protection, namely authorship and originality;[33] which in turn require human attributes[34]. For instance, the Software Directive and the Database Directive define authorship on the basis of the natural person(s) or group(s) of natural persons who created the work[35]. The Preamble to the Copyright Term Directive defines an original work as the “author’s own intellectual creation reflecting his personality.”[36] Several other EU directives also link originality to natural persons or human attributes[37].
The Court of Justice of the European Union has also decided several landmark cases on copyright protected works[38]. In these cases, the Court held that copyright only attached to original works; and in determining whether originality existed, the Court used references such as “author’s own intellectual creation”, “the author’s personality”, or the “author’s personal touch.”[39] In Infopac[40], the Court had to determine whether originality subsisted in news clippings that were created through a data capture process (where images of the original articles were scanned, then translated into text, and a 11-word snippet was created). The Court noted that the elements of the clippings by themselves may not have originality, but the selection process could warrant originality; and ruled that “regarding the elements of such works covered by the protection, it should be observed that they consist of words which, considered in isolation, are not as such an intellectual creation of the author who employs them. It is only through the choice, sequence and combination of those words that the author may express his creativity in an original manner and achieve a result which is an intellectual creation.”[41]
In the circumstances, whether AI generated work will have copyright protection under EU laws may depend on how much input comes from the programmer, and how much from the AI[42]. For instance, in The Next Rembrandt Project, it appears that a lot of work was performed by the team of computer programmers and art experts in selecting the data to feed the AI, and setting the parameters for the portraits to be created. Arguably, the personality of the authors could be represented through their careful curation of the data selected to feed the AI[43]. The art works created by this AI might therefore have the requisite human attributes to attract copyright protection.
Conversely, where the selection process is more automated and undertaken by the AI, it would probably lack the human attributes required under the current EU case law. Most EU scholars have concurred that these types of AI generated works probably cannot be considered original works and will not be protected by copyright[44].
Moving forward, the EU’s Committee on Legal Affairs has made recommendations to the Commission on Civil Law Rules on Robotics[45]. They recommend the elaboration of the criteria for “own intellectual creation” for copyrightable works produced by computers or robots[46]. They have also suggested granting legal status (including specific rights and obligations) to the most sophisticated autonomous robots[47]. This is still a developing area under EU law, and it would be interesting to see whether these recommendations translate to granting the AI copyright in the works they have generated.
The UK Copyright, Designs and Patents Act 1988 (the “CDPA”) refers to “computer-generated work”[48] as work generated by a computer in circumstances where there is no human author. The CDPA then provides that “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.”[49] The duration of such copyright protection is 50 years from the end of the calendar in which the work was made[50].
The only UK case applying s9(3) of the CDPA is Nova Productions Ltd v Mazooma Games Ltd[51]. In this case, the Court had to decide whether individual frames shown on a screen when playing a computer game were “computer-generated” works authored by the programmer. The court held that the programmer was deemed to be the author, because he “devised the appearance of the various elements of the game and the rules and logic by which each frame is generated and he wrote the relevant computer program”, and was therefore “the person by whom the arrangements necessary for the creation of the works were undertaken”[52].
Interestingly, the Court considered the role of the user. The Court acknowledged that the appearance of the graphics and frames could depend on how the game was played by the user. However, the user is not the author of the artistic works created in the graphics and frames, because “his input is not artistic in nature and he has contributed no skill or labour of an artistic kind. Nor has he undertaken any of the arrangements necessary for the creation of the frame images. All he has done is to play the game.”[53]
Based on the foregoing, it seems that the CDPA could provide copyright protection for AI generated works. Such copyright protection would be given to the person who made the necessary arrangements to create the work[54] - who could be either the programmer or possibly even the user who has contributed skill and labour of an artistic kind[55]. However, observers have opined that this interpretation could be problematic, especially in the light of AI generated works, because it is uncertain who makes the “necessary arrangements” (is it the programmer or the user?).[56] Furthermore, a low threshold for the requirement to make “the necessary arrangements” could lead to a virtual monopoly in AI generated works[57]. Additionally, the requirement for originality is unclear[58].
This essay contends that despite the foregoing concerns, there are advantages to the UK’s approach, and it is better than the current position under Australian and EU laws, which likely consider AI generated works as not worthy of protection. Part III below explores the general policy reasons for granting copyright protection, while Part IV considers how the UK approach might be adapted to address the concerns in the paragraph above, and form a basis for reforming Australia’s copyright laws.
Having observed that Australian copyright laws are unlikely to protect AI generated works, this Part contends that Australia should attribute copyright to AI generated works for the following reasons:
The digital age has brought about advances in technology and an exponential growth of investments in AI. Market research estimates that investments in AI will rise from $640 million in 2016 to $37 billion by 2025[59], and the revenue generated from the direct and indirect application of software based on AI will grow from $1.4 billion in 2016 to $59.8 billion by 2025[60].
There will be diverse applications of AI and machine learning, and AI will undoubtedly be more involved in the creation of music, works of art, and novels. The examples of the creative works generated by AI (see Part I above) illustrate how far we have come from AI assisted works to AI generated works that are creative and profit generating. There is therefore an urgent need to effectively regulate the production and use of such AI generated works[61].
The fundamental purpose of copyright law is to reward and incentivise creativity and innovation[62]. Arguably, awarding copyright protection would encourage AI programmers to invest their time and skills into the AI sector, and stimulate financial investments from companies, which are both important contributors to the research and development of the AI sector.[63]
Given that the Australia has launched its Tech Future initiative and aims to maximise opportunities for the digital economy by stimulating investments and incentivising innovation[64], it is timely for Australia to consider revising its copyright laws for AI generated works.
Admittedly, there are concerns that granting copyright could lead to corporate developers having a virtual monopoly over the myriad of works generated by the AI[65]. But this is where copyright serves as a useful regime by which to control the developer’s monopoly, regulate the use of the works, and ensure public access to the works[66]. In the absence of a definable copyright regime, there would be no effective way to balance the control and use of the AI generated works amongst the investors, programmers and end users, which could exacerbate concerns about any party having a monopoly.
Attributing copyright subsistence also provides certainty of ownership. Certainty of ownership mitigates wasted expenses and time over potential litigation suits, as ownership is less likely to be contested[67].
Some observers have suggested that other legal arrangements instead of copyright, such as contractual relationships and Creative Commons licences, could be utilised to allocate use and control rights over AI generated works. However, these other legal arrangements would only make sense if the base property in AI generated works is copyright[68]. Contractual rights and licences would exist in a vacuum if there is no base property.
In any event, relying on such other legal arrangements could be problematic. Having other legal arrangements outside the domain of copyright protection could result in substantially identical subject matter being subject to both copyright laws and the rules of the legal arrangement, which creates areas of uncertainty[69]. Therefore, it is submitted that to ensure clarity and consistency in the law, AI generate works should only fall under the domain of copyright law.
Other common law countries such as New Zealand, India, Ireland, Hong Kong and South Africa have adopted the same wording from the UK’s CDPA s 9(3). Given that these common law countries recognise that copyright can be attributed to AI generated works, it seems that Australia is out of step with its counterparts. There appears to be no strong policy reason as to why Australia should not adapt its laws too[70]. In fact, by harmonising its laws with these jurisdictions, it may encourage AI related trade and investment in Australia[71].
Some scholars have argued that AI should be granted legal personality, and that copyright authorship and ownership should accordingly be vested in the AI[72]. Conceivably, this is a novel way to encourage the growth and development of AI. However, this essay does not agree that AI should be vested with copyright authorship and ownership, permitting this would be highly controversial and would lead to more questions instead of answers – there may be ethical issues[73], questions as to whether the AI is legally responsible in a court of law[74], and how the AI would exploit its copyrights.
This essay contends that there is little reason to place copyright law into such turmoil. Instead of revolutionary transformation, copyright law should build on and extend existing laws to address technological change[75]. In this regard, the essay suggests that based on the reasons above, copyright law should adapt to afford copyright protection to AI generated works; and vest such copyright in the persons who made the necessary arrangements for the creation of the work. Part IV below explores how Australian copyright law may apply these suggested adaptations.
These suggested reforms are not intended to be exhaustive, but merely to generate discussion.
As illustrated in Part II(A) above, the obvious issue was that Australian laws do not recognise AI generated works as authored works. Therefore, one suggestion is for the CA to recognise a category of “AI generated work”, where such work is generated by AI in circumstances where there is no identifiable human author.
The definition from s 178 of the CDPA could be adopted. However, in view that AI generated works are typically only contentious when it is difficult to identify a human author or find clear evidence of human authorship in the AI generated works, the definition from s 178 of the CPDA could be revised slightly to read as follows – “AI generated”, in relation to a work, “means that the work is generated by AI in circumstances such that there is no clearly identifiable human author of the work”[76]. AI generated works could be retained in Part III of the CA as “works”.
The CA could adopt a provision similar to s 9(3) of the CDPA, and provide that “in the case of a literary, dramatic, musical or artistic work which is AI generated, the author shall be taken to be the person by whom the arrangements necessary for the creation of the work are undertaken.” This approach recognises that authorship of such AI generated works can be attributed to the person who made the necessary arrangements for the creation of the work.
As mentioned in Part II(C) above, some concerns about this approach are that it remains ambiguous who the author is, as there are many parties involved in the creation of the work (such as the programmer of the AI, the person who feeds data to the AI, or the end user).[77] However, such ambiguity can be solved simply by applying the rule on a case by case basis[78]. This apparent ambiguity is actually a necessary flexibility, because AI generated works are not homogenous, and authorship needs to be determined on a case by case basis[79].
One of the concerns raised in Part II(C) was that a low threshold for the requirement to make “the necessary arrangements” could lead to authors having a virtual monopoly over the AI generated works. The identification of a person as the author will turn on the arrangements which the person made.
In considering who “is the person by whom the arrangements necessary for the creation of the work are undertaken”, Australian case law on determining the maker of a film could be instructive. The CA s 22(4)(b) provides that the maker of a film “is the person by whom the arrangements necessary for the making of the film were undertaken”– these are analogous to our proposed wordings for AI generated works. In considering s 22(4)(b) of the CA, the Court in Seven Network (Operations) Ltd v TCN Channel Nine Pty Ltd [80] held that the relevant person is generally the producer who makes the financial or administrative arrangements for the production of the first copy of the film.[81]
The UK’s CPDA also contains similar provisions to determine the producer of a sound recording or film.[82] UK case law on these provisions generally take into account the following key factors in determining who the relevant person is: (i) who intended to make the film; (ii) who paid for the making of the film; and (iii) whether there would have been a film if not for the initiative and conduct of that relevant person[83].
The above cases could be applicable to AI generated works. Extrapolating on the above case law, this essay suggests that the following factors could be used to determine whether the arrangements were necessary for the creation of the AI generated work, and the identity of the author:
• Who instigated the work with the intention of creating it[84].
• While the person who started the process is important, there should also be proximity to the final output[85]. In a scenario where a programmer creates an AI software, which is then run by any user to generate music (such as in the Flow Machines), it would seem counterintuitive that the programmer should own all music generated by the AI. After all, the user is the one setting the parameters and the genre for the music to be created.
• The extent to which the arrangements contribute to the final output of the work[86]; greater weightage should be given to the arrangements which are more crucial to the materialisation of the work. In this regard, merely pressing a button and letting the AI do all the work would arguably not make any material contribution to the final output.
This essay submits that these factors could mitigate the concerns over expanding copyright to include an exponentially greater array of protected works.
As mentioned in Part II(C) above, one concern was how originality in the AI generated works was to be determined. Based on Australian case law (see Part II(A)), originality requires independent intellectual effort.
To address this issue, it would be relevant to consider the process for creating AI generated works. Guadamuz points out that it is often not “just a matter of pressing a button and letting the machine do all the work”.[87] As it was in the case of the Next Rembrandt Project, the art experts and computer engineers had to programme and teach the AI to paint; curate the data input to the AI, and set the parameters for the portraits to be created. Guadamuz argues that the process requires human input which is “both lengthy and full of intellectual creativity”[88] and suggests that this could “potentially be considered as carrying enough intellectual creation to the AI generated work”.[89] This is an interesting approach which looks at the process as the focal point of the originality test, instead of placing the output in the central focus of the originality test.
Admittedly, in IceTV and the Phone Directories cases, the court focused on the originality in the output (i.e. the work itself) and required that independent intellectual effort be taken in the fixing of the work in its material form. Adopting an approach which applies the originality test to the process of authorship, instead of the act of fixation might run counter to the ruling in these cases. However, copyright law already has the capacity to house such contradictions. This is apparent in the case of photography. The invention of photography had the effect of extending authorship to instances where the act of fixation is done by a machine[90]. By extrapolation, the originality test is applied during the process of taking the photo. In an EU case concerning the portrait photograph of a school girl taken by an Austrian photographer[91], the Court held that the photographer was able to “express his creative abilities in the production of the work by making free and creative choices” thereby stamping the work with his “personal touch”. If work could be thought to be original if there is enough human input in the process that produced the work, then this principle could also be extended to AI generated works[92].
As technology develops, we are likely to head in the direction where more creative works are autonomously created by AI. This raises issues about the protectability of these AI generated works.
There is a clearly a need to regulate the production and use of these works, and copyright has the potential capacity to extend its framework to address the challenges brought about by AI generated works.
There could be other legal solutions beyond the realm of copyright, and we cannot predict which options Australia or the EU may eventually take. But one thing is for certain, there is merit in acting sooner rather than later, and we cannot ignore the rise of the machines. This haiku created by Ray Kurzweil’s Cybernetic Poet (an AI) may be haunting in the context of this essay:
“The stifling stuffy
Catholic schoolroom,
where I cannot be real.”[93]
* I would like to thank Dr Catherine Bond for her guidance and support in developing this paper.
1 Andres Guadamuz, ‘Artificial intelligence and copyright’, (2017) 5 WIPO Magazine 14, 17.
[2] Niloufer Selvadurai and Rita Matulionyte, ‘Reconsidering creativity: copyright protection for works generated using artificial intelligence’ (2020) 15(7) Journal of Intellectual Property Law and Practice 536.
[3] Ibid.
[4] Madeleine de Cock Buning, ‘Autonomous Intelligent Systems as Creative Agents under the EU framework for Intellectual Property’ (2016) 7(2) European Journal of Risk Regulation 310, 312 – 313.
[5] Ibid.
[6] Selvadurai and Matulionyte (n 2) 539.
[7] Kalin Hristov, ‘Artificial Intelligence and the Copyright Dilemma’ (2017) 57(3) IDEA: The Journal of the Franklin Pierce Centre for Intellectual Property 431, 434.
[8] The Next Rembrandt (Web Page) <http://www.nextrembrandt.com> .
[9] Ahmed Elgammal, ‘What the Art World Is Failing to Grasp about Christie’s AI Portrait Coup’, Artsy (Web Page, 29 October 2018) <https://www.artsy.net/article/artsy-editorial-art-failing-grasp-christies-ai-portrait-coup >; see also Gabe Cohn, ‘AI Art at Christie’s sells for $425,500’, The New York Times (Web Page, 25 October 2018) <https://www.nytimes.com/2018/10/25/arts/design/ai-art-sold-christies.html>.
[10] Danny Lewis, ‘An AI-Written Novella Almost Won a Literary Prize’, Smithsonian Magazine (Web Page, 28 March 2016) <https://www.smithsonianmag.com/smart-news/ai-written-novella-almost-won-literary-prize-180958577/ >.
[11] ‘AI makes pop music’, Flow Machines (Web page, 19 September 2016) <https://www.flow-machines.com/history/events/ai-makes-pop-music/ >.
[12] ‘Musical Creativity Machines’, Imagination Engines Inc (Web Page) <http://imagination-engines.com/iei_musical_composition.php> .
[13] Massimo Maggiore, ‘Artificial Intelligence, computer generated works and copyright’ in Enrico Bonadio and Nicola Lucchi (eds), Non-Conventional Copyright: Do New and Atypical Works Deserve Protection? (Edward Elgar Publishing, 2018) 382, 383.
[14] Jani McCutcheon, ‘The Vanishing Author in Computer Generated Works: A Critical Analysis of Recent Australian Case Law’ (2013) 36 Melbourne University Law Review 915, 929 (‘The Vanishing Author’).
[15] Pratap Devarapalli, ‘Machine Learning to Machine Owning: Redefining the Copyright Ownership from the Perspective of Australian, US, UK and EU Law’ (2018) 40(11) European Intellectual Property Review 722.
[16] Andres Guadamuz, ‘Do Androids Dream of Electric Copyright? Comparative Analysis of Originality in Artificial Intelligence Generated Works’ (2017) 2 Intellectual Property Quarterly 169, 175 (‘Do Androids Dream of Electric Copyright?’).
[17] CA s 35.
[18] Desktop Marketing Systems v Telstra Corp [2002] FCAFC 112 (‘Desktop Marketing’).
[19] IceTV Pty Limited v Nine Network Australia Pty Limited [2009] 239 CLR 458 (‘IceTV’).
[20] Ibid [54].
[21] Ibid [33] - [34].
[22] Ibid [95] - [106].
[23] Cameron Andrews, ‘Copyright in computer-generated work in Australia post-IceTV: Time for the Commonwealth to Act’ (2011) 22 Australian Intellectual Property Journal 29, 38.
[24] Telstra Corp Ltd v Phone Directories Co Pty Ltd [2010] FCAFC 149; [2010] 273 ALR 725 (“Phone Directories”).
[25] Ibid [113].
[26] Ibid [114].
[27] Ibid [117] – [118].
[28] Acohs Pty Ltd v Ucorp Pty Ltd [2010] FCA 577 (“Acohs”).
[29] Ibid [50].
[30] Ibid [53].
[31] McCutcheon, ‘The Vanishing Author’ (n 14) 931.
[32] Guadamuz, ‘Do Androids Dream of Electric Copyright?‘ (n 16) 178.
[33] Ibid.
[34] Bob L. T. Sturm et al, ‘Artificial Intelligence and Music: Open Questions of Copyright Law and Engineering Praxis’ (2019) 8(3) Arts, Multidisciplinary Digital Publishing Institute 115, 118 (‘Artificial Intelligence and Music’).
[35] Maria Iglesias, Sharon Shamuilia and Amanda Anderberg, Publications Office of the European Union, Intellectual Property and Artificial Intelligence - A Literature Review (EUR 30017 EN, 2019), 14 <https://publications.jrc.ec.europa.eu/repository/handle/111111111/58660 > (“EU Report”).
[36] Guadamuz, ‘Do Androids Dream of Electric Copyright?‘ (n 16) 178.
[37] EU Report (n 35) 14.
[38] EU Report (n 35) 14; See also Infopaq International A/S v Danske Dagbaldes Forening (C-5/08, EU:C:2009:465, 16 July 2009)(“Infopaq”); and Football Dataco Ltd and Others v Yahoo! UK Ltd and Others (C- 6-604/10, ECLI:EU:C:2012:115, 1 March 2012)(“Dataco case”).
[39] Sturm et al, ‘Artificial Intelligence and Music’ (n 34) 118; See also Dataco case (n 38) [38].
[40] See Infopac (n 38).
[41] Ibid [45].
[42] Guadamuz, ‘Do Androids Dream of Electric Copyright?‘ (n 16) 179.
[43] Ibid.
[44] EU Report (n 35) 14.
[45] Committee on Legal Affairs, European Parliament, Report with recommendations to the Commission on Civil Law Rules on Robotics (2015/2103(INL), 27 January 2017) <https://www.europarl.europa.eu/doceo/document/A-8-2017-0005_EN.html#:~:text=1).&text=(1)%20A%20robot%20may%20not,conflict%20with%20the%20First%20Law.>.
[46] Ibid Explanatory Statement.
[47] Ibid [59(f)].
[48] CDPA s 178.
[49] CDPA s 9(3).
[50] CDPA s 12.7.
[51] Nova Productions Ltd v Mazooma Games Ltd [2007] EWCA Civ 219 (“Nova Productions”).
[52] Ibid [105].
[53] Ibid [106].
[54] Devarapalli (n 15) 724.
[55] Guadamuz, ‘Do Androids Dream of Electric Copyright?‘ (n 16) 177.
[56] Anne Lauber-Ronsberg and Sven Hetmank, ‘The concept of authorship and inventorship under pressure: Does artificial intelligence shift paradigms?’ (2019) 14(7) Journal of Intellectual Property Law & Practice 570, 574.
[57] Jani Ihalainen, ‘Computer creativity: artificial intelligence and copyright’ (2018) 13(9) Journal of Intellectual Property Law & Practice 724, 725.
[58] Lauber-Ronsberg and Hetmank (n 56) 574 – 575.
[59] Alastair Peet and Tom Wilde, ‘Artificial intelligence: the investment of 2017 and beyond’, Financier Worldwide (Web Page, February 2017) <https://www.financierworldwide.com/artificial-intelligence-the-investment-of-2017-and-beyond#.X4_-PRLiuUk>.
[60] Business Wire, ‘Artificial Intelligence Software Revenue to Reach $59.8 Billion Worldwide by 2025, According to Tractica’ (Web Page, 2 May 2017) < https://tractica.omdia.com/newsroom/press-releases/artificial-intelligence-software-revenue-to-reach-59-8-billion-worldwide-by-2025/>.
[61] McCutcheon, ‘The Vanishing Author’ (n 14) 956.
[62] Ibid 951.
[63] Hristov (n 7) 444; See also McCutcheon, ‘The Vanishing Author’ (n 14) 954.
[64] Australian Government, Department of Industry, Science, Energy and Resources, ‘Australia’s Tech Future’ (Web Page, December 2018) <https://www.industry.gov.au/data-and-publications/australias-tech-future/implementing-the-strategy>.
[65] Ihalainen (n 57) 725.
[66] McCutcheon, “The Vanishing Author” (n 14) 956.
[67] Ibid.
[68] Dilan Thampapillai, “If Value Then Right? Copyright and Works of Non-human Authorship” (2019) 30 Australian Intellectual Property Journal 96, 98.
[69] McCutcheon (n 14) 966.
[70] Ibid 956 - 957.
[71] Ibid.
[72] Hristov (n 7) 441; See for e.g. R Pearlman, ‘Recognizing Artificial Intelligence (AI) as Authors and Inventors Under U.S. Intellectual Property Law’ (2018) 24(2) Richmond Journal of Law and Technology 1, 30.
[73] Devarapalli (n 15) 727.
[74] Hristov (n 7) 441.
[75] Selvadurai and Matulionyte (n 2) 538.
[76] McCutcheon, ‘The Vanishing Author’ (n 14) 959.
[77] Lauber-Ronsberg and Hetmank (n 56), 574.
[78] Andes Guadamuz, ‘Impact of Artificial Intelligence on IP Policy’ (Submission to WIPO, Draft Issues Paper on Intellectual Property Policy and Artificial Intelligence, WIPO/IP/AI/2/GE/20/1, 13 December 2019) <https://www.wipo.int/export/sites/www/about-ip/en/artificial_intelligence/call_for_comments/pdf/ind_guadamuz.pdf > (“WIPO Submission”).
[79] Jani McCutcheon, ‘Curing the Authorless Void: Protecting Computer-Generated Works Following IceTV and Phone Directories’ [2013] MelbULawRw 9; (2013) 37(1) Melbourne University Law Review 46, 54 (‘Curing the Authorless Void’).
[80] Seven Network (Operations) Ltd v TCN Channel Nine Pty Ltd [2005] 146 FCR 183.
[81] Ibid at 186 -188.
[82] CPDA s 178.
[83] McCutcheon, ‘Curing the Authorless Void’ (n 79) 64.
[84] Ibid 55 – 56.
[85] Ibid.
[86] Ibid.
[87] WIPO Submission (n 78).
[88] Ibid.
[89] Ibid.
[90] Thampapillai (n 68) 112.
[91] Eva-Maria Painer v Standard VerlagsGmbH and Others (C-145/10, ECLI:EU:C:2011:798, 1 December 2011).
[92] WIPO Submission (n 78).
[93] Ray Kurzweil, The Age of Spiritual Machines: How We Will Live, Work and Think in the New Age of Intelligent Machines (Texere Publishing, 2001), 119.
AustLII:
Copyright Policy
|
Disclaimers
|
Privacy Policy
|
Feedback
URL: http://www.austlii.edu.au/au/journals/UNSWLawJlStuS/2020/26.html