Introduction
“Once we start seeing ourselves only in terms of data, we have reduced ourselves to something that can be bought and sold. This is particularly relevant in the ongoing debates about large language models (LLMs). When you write down a thought, it is tempting to think that your thought and the language data you create are the same…. Our minds produce outputs and data, but we are more than just what we generate. Just like we work to preserve the intangible cultural heritage of societies, we should also protect the intangible aspects of our mental worlds.”1
— Dr Ophelia Deroy
It is a well-established rule in the world of “intellectual property rights” that copyright cannot be provided for an “idea” but for the “expression” of that idea in its various forms. Time and again through various precedents, it is also well-established that this “idea-expression dichotomy” is sorted out by balancing the “sweat of brow” and “modicum of creativity” with all its nuances. This traditional rulebook of anthropocentric copyright laws faced a big blow with the invention of “artificial intelligence” and the boom of digital world. Suddenly the traditional copyright laws started appearing as means of colonial subjugation hampering the juggernaut of contemporary generative-artificial intelligence (AI) based creations and the technocentric metaverse made space for new laws to be created to cater the new age disputes. Adding turbidity to the murky waters of this digital conundrum came the new age brain-computer interface (BCI)-based neurotechnologies to obliviate the grey zone of idea and expression and opened up the “pandora’s box” where “idea” remains no longer distinctly separated from the expression through “BCI”. If articulation of “A Brief History of Time” by famous scientist Dr Stephen Hawking is an example of triumph of science over grave debilitating disease like amyotrophic lateral sclerosis (ALS), it is also an example of “BCI”-based creation where the “sweat of brow” is taken over by machine to express “ideas” from brain, leaving “modicum of creativity” at the discretion of “science”, creating novel copyright dilemmas for a digital metaverse while exploring the history of the universe. This dilemma goes beyond the issue of “fair dealing” of using copyrighted literature by scholars, or beyond the much-debated issue of original creator’s intellectual property right (IPR) violations while feeding “machine learning models” of “generative AI” upon those data for new creations and hits the very core of the copyright laws: whether “idea” can be copyrighted when creator is a machine? Whether the “sweat of brow” doctrine a dead concept? Whether the “modicum of creativity” can be attributed to a machine? Whether a machine can be provided with “legal personhood” making them able to hold the copyright of a creation? And finally, how to solve the riddle of legal accountability of “BCI” based AI-generated content in case of IPR infringement disputes? This article is an attempt to navigate through the conundrum of all these new age issues, exploring various facets of copyright dilemmas of a technocentric metaverse using the age-old anthropocentric copyright laws and precedents as the beacon.
New age tools: BCI and generative AI
Brain-computer interface (BCI): Brainchild of Jacques Vidal, electroencephalography (EEG) based BCI was first introduced to the world in 1973.2 BCI is a mode of direct brain to machine interaction bypassing the physical body where a person can do things just by thinking of it and articulate new creations just by developing the idea in the mind. BCI interface contains sensors which detect the thought of a person through electrical signals of the brain and using AI based pattern recognising algorithms convert them into commands to be executed by the machine. BCI device can write poems, stories, novels or draw pictures as per the ideas of a person’s brain which in case of a normal person is done through his own physical hands.
Generative artificial intelligence (AI): Father of “artificial intelligence” John McCarthy defined AI as “… making a machine behave in ways that would be called intelligent if a human were so behaving”. In pursuit of advancing this science of making the machines smarter, novel ideas and novel dilemmas have taken birth. From “Turing test” by Dr Alan Turing to evaluate “machine intelligence”3 to Dr Shlomit Yanisky Ravid’s discourse in cognitive abilities of AI systems, numerous instances have made it clear that AI systems are no longer in their nascent stage but have grown up enough to demand legal personhood and face liability issues.4 Generative AI based machine learning models learn from the existing data and recreate new data as per need. “LLM” based chatbots like “ChatGPT” or text to image generator like “Midjourney” or “DALL-E” and text to video generator like “Sora” are commonly in use. Lately, a plethora of AI based apps have come up and being used by millions of people to create contents overwhelming the internet servers and raising serious copyright issues.
Deciphering “idea-expression dichotomy” at the deathbed of “sweat of brow doctrine”
Before analysing this topic lets go back to nineteenth century when cognitive psychology and neuroscience was still at its immature stage and neurotechnology was unheard of. In a time when “idea” of mind was still veiled behind the fog of ignorance, the US Supreme Court faced a novel dilemma to decide “whether a photograph can be copyrighted” in “Burrow-Giles Lithographic Co. v. Sarony”5, considering the fact that it is “not written or created by an author (literary production)” but a production of photography machine (camera). Interestingly, the riddle was solved by the wise Judges holding “photograph” as a “visible expression” of “ideas in mind of author”. Almost a century and half later in “Naruto v. David John Slater”6, well-known as “The Monkey Selfies case”, when the same issue of photography was created by a monkey instead of human, it was held, “the Copyright Office will not register works produced by ‘nature, animals or plants’”, as that “Supreme Court and Ninth Circuit have repeatedly referred to ‘persons’ or ‘human beings’ when analysing authorship under the Copyright Act, 19767.” The “Compendium of US Copyright Office Practices” specifically mentions, “to qualify as a work of ‘authorship’ a work must be created by a human being”. So, what happens if the “expression” of an idea is not physically created by the “human author” but instead created by “mind” using “artificial intelligence” which is fed upon a million other creations? Who would be recognised as the author of such creations? And who would be held liable for the copyright breach of other creators when the expression is generated by such “non-human” entities? Just like the eighteenth century “Burrow-Giles Lithographic case”8, standing at this liminal age of digital transformation, we are facing another novel dilemma, “Can there be a copyright for the photograph or image deciphered from human brain using BCI?”
In disputes regarding copyright, while the doctrine of “sweat of the brow”9 had been developed through case precedents like “Walter case”10 and “Cummins case”11, the concept of “modicum of creativity”12 was highlighted in cases like “Eastern Book Co. case”13 and tests like “Lay observer’s test” was resorted to in “R.G. Anand v. Delux Films”14 to solve the dilemma of “idea-expression dichotomy” in “substantial similarity in copyright infringement”. In “Shamoil Ahmad Khan v. Falguni Shah”15 Bombay High Court developed “extraction test” stating that “seed of an idea grows into a theme”. No doubt these were only continuation of precedents that started long back in “Baker v. Selden”16 and evolved in “Mazer v. Stein”17 delineating “idea” as different from “expression”. But as science advanced, BCI started deciphering ideas of brain, and AI took charge of expressing those ideas. In the “metaverse” of “virtual reality” computer programmes became the architects of “scènes à faire”18 and gradually the importance of the concept of “sweat of brow” for expressing an idea which gave weight to the demands of IPR lost its relevance.19
What can be copyrighted?
“The Copyright Act, 1957”20 of India in “Section 13” mentions the creations that can be copyrighted as: “copyright shall subsist throughout India in the following classes of works: (a) original literary, dramatic, musical and artistic works; (b) cinematograph films; and (c) sound recording.” On the other hand the US “Copyright Act of 1976” provides “copyright protection” to “original works of authorship fixed in any tangible medium of expression, now known or later developed, from which they can be perceived, reproduced or otherwise communicated, either directly or with the aid of a machine or device,”21 and contains four vaguely defined criteria: (i) “work of authorship”; (ii) “originality”; (iii) “fixation”; and (iv) “idea-expression dichotomy”.
Who is an author?
Section 2(d) of the Copyright Act, 1957 while defining “author” mentions various forms of “human author” be it a “composer”, “artist”, “photographer” or a “film producer”. Nowhere in the Act any provision has been kept to provide “AI” any kind of recognition as an “author”. Though Section 2(d)(vi) of the Copyright Act, 1957 mentions that “person creating computer generated work” is the author of the same. That means “generative AI”, even if generates a creation itself never gets any recognition as per the Copyright Act, 1957 unless a human claims authorship for the creation. Certainly, mention of “granting copyright protection” for the period of “lifetime of author plus fifty years after death” in “The Berne Convention”22 points towards a mortal “human” author and not an immortal machine. UK’s “Copyright, Designs and Patents Act, 1988 (GB) (CDPA)”23 provides copyright protection to the AI generated work but does not recognise AI as author. Rather the person who arranges for the creation through AI is given copyright. US copyright laws throughout the history have been strong supporter of human author. Though in “Goldstein v. California”24, it held that definition of “creator” or “author” “have not been construed in their narrow literal sense but, rather, with the reach necessary to reflect the broad scope of constitutional principles”,25 yet in practicality all application for copyright protection of AI generated content has met with similar fate in the US copyright office. Refusal of copyright protection to “A Recent Entrance to Paradise”, an artwork created by AI system “Creativity Machine” (developed by Dr Stephen L. Thaler) or providing partial and limited copyright to graphic novel “Zarya of the Dawn” (Author Kris Kashtanova) having images generated by “AI system” “Midjourney” shows the importance given to “human involvement” by the US copyright laws. Perhaps South Africa’s “Companies and Intellectual Property Commission (CIPC)” was the first office to grant authorship to an AI system when it recognised Device for the Autonomous Bootstrapping of Unified Sentience (DABUS) of Dr Thaler as an author in July 2021. In 2020 Indian copyright office for the first-time had recognised AI based painting app Robust Artificially Intelligent Graphics and Art Visualiser (RAGHAV) as a co-author of artwork “SURYAST” developed by feeding the AI system with Vincent Van Gogh’s art style, though sole authorship of the AI system without human author (Ankit Sahni) was rejected. Withdrawal notice of the same came from the copyright office a year later, but as a positive outcome the “Parliamentary Standing Committee” recommended to review existing copyright laws to cater the AI and AI related inventions. Still when it came to actions, to everybody’s surprise, the Press Release by “Ministry of Commerce & Industry” in its statement in 9-2-2024 denied the need to create any separate rights for “AI generated works”.
The dilemma of “idea-expression dichotomy”
Article 9(2) of the Agreement on Trade-Related Aspects of Intellectual Property Rights (TRIPS), 1995 (TRIPS Agreement)26 or Article 2 of the “WIPO Copyright Treaty (WCT), 1996”27 resonate the same perception as the “Guidance Notes” on the “Berne Convention” and so does the landmark judgments of famous cases like “Donoghue v. Allied Newspapers Ltd.”28 (held that “there is no copyright of an idea”29 and “ideas can be communicated but not copyrighted”) or “Barbara Taylor Bradford v. Sahara Media Entertainment Ltd.”30, (held that “Copyright protected originality of expression and not the originality of idea behind it”31), all establishing copyright to be limited to “expression” and not “idea”.
Taking cue from the “Burrow-Giles Lithographic case”32, one might ask, if an image generated by the “hardware” called “camera” can be copyrighted then why cannot an image generated by a “software” named “Midjourney” or “Creativity Machine” be copyright protected? Going to our main focus, question arises, what would happen if the image were generated by the “BCI” from the mind of the author. Certainly, as per present copyright laws which were made before the BCI-era, it would be considered mere an idea and machine being non-human entity cannot be an author. So, who would be regarded as the author if ideas cannot be copyrighted? Who is giving expression to the idea? Certainly, it is the “BCI” device with AI based algorithms which is expressing that idea in art form. Will then a disabled person who cannot express the ideas of his brain through his amputated physical limbs and instead expressed them through BCI-devices using AI based algorithms be deprived of copyright of his own ideas expressed with assistance of machines? With the advancement of neurotechnology, the question of “authorship” converts into the dilemma of “idea-expression dichotomy” as the gap between “idea” and “expression” gets obliviated in the technocentric future.
Patenting “concept” as abstract “Idea” gets material form
As science deciphers “idea” into “electrical activities” of cerebral cortex and “BCI” based AI algorithms give material form to those potential differences, AI based mind-controlled devices narrow down the distance between “idea’s and expression’s” burying the “sweat of brow doctrine” and becomes the medium of realising the “modicum of creativity”. What to copyright then in such a situation? As we move towards the metaverse of “virtual reality”, copyrighting the “concept” would be the wise middle path taken to solve the “idea-expression dilemma”.
Other copyright issues of BCI-AI interface generated creations
“Thought broadcasting rights” in BCI era: As thought broadcasting becomes a reality with BCI-AI interface, and communicating through BCI-devices without speech-based telecommunication brings hope to patients of neurological diseases and as storing of brain-scanned data in the cloud and broadcasting that data no longer remains an impossible task, novel issues of broadcasting rights come into picture. “The Copyright Act, 1957” under Chapter VIII provides copyrights to “broadcasting organisation” and Section 3733 under this chapter specifically mentions “broadcast reproduction rights”. But the question remains, who gets the broadcasting-rights of “brain data”? Is it the BCI device that is broadcasting brain data, is it the third person creating the BCI device or is it the brain itself which is producing the data?
Copyright issues of brain-net: Just like connecting multiple computers through intra-net, it is now possible to connect multiple brains with BCI-AI interface devices. What happens to the copyright of the creation of that brain-net with conjoint effort of multiple BCI devices is a novel question hard to answer with archaic copyright laws.
Fair use, rights of AI models for learning on datasets and legal accountability of infringements
In 1841 J. Joseph Story’s judgment in “Folsom v. Marsh”34 paved the way for establishing “four-factor analysis” used to determine “fair use” in US common law.
The four factors taken into consideration were:
(a) “What is the purpose and character of the use?
(b) What is the nature of the copyrighted work used?
(c) What is the amount and substantiality of the original work used?
(d) What is the effect upon original work’s value?”
Finally, the provision of “fair use” got incorporated into the “Copyright Act of 1976” Section 10735. Conventionally, the Supreme Court of USA had categorised the provision of “fair use” as a sort of “affirmative defence”, as in “Campbell v. Acuff-Rose Music, Inc.”36 In other words, in cases of “copyright infringement”, the defendant had to bear the burden of proving that it was a “fair use” and “not an infringement”. But in “Stephanie Lenz v. Universal Music Corpn.”37, it was held that “fair use” is not merely a “defence to an infringement claim”, but it is an “expressly authorised right”. It also stated that it can be considered “an exception” to the “exclusive rights” granted to the original creator by copyright law. Section 107 of the US Copyright Act, 1976 states that fair use for purpose of — “criticism, comment, news reporting, teaching (including multiple copies for classroom use), scholarship, or research” is not an “infringement of copyright”. The Berne Convention did not mention provisions such as “fair use” or “fair dealing”, but it churned out the “Three-Step Test” in Article 9(2), which articulated a rough framework for all the World Trade Organisation (WTO) nations to make their own “national exceptions for fair use”. It says original work can be used in: “(i) special cases; (ii) the original work must not be exploited by the reproductive work; (iii) the new work must not be prejudicial against the interest of the original author”. “The Copyright Act, 1957” has provisions for allowance of reproduction of certain copyrighted material without the permission of the original copyright holder. Section 5238 of the “Copyright Act, 1957” states: “Certain acts not to be infringement of copyright.” Section 52(1) enlists the “Acts which shall not constitute an infringement of copyright.” The Copyright Act, 1957, Section 52(1)(a) mentions “fair dealing.” It states:
52. Certain acts not to be infringement of copyright.—(1)(a) The following acts shall not constitute an infringement of copyright, namely: (a) a fair dealing with any work, not being a computer program for the purpose of: (i) private or personal use, including research; (ii) criticism or review, whether of that work or of any other work; (iii) the reporting of current events and current affairs, including the reporting of a lecture delivered in public.
Whether training of AI models on pre-existing literature or artworks is fair use needs to be analysed on case-to-case basis, but the present copyright laws of India remain vague in this matter. Australia and UK provide certain exceptions for “text and data mining” by AI models for common good. The Digital Millennium Copyright Act, 1998 (US)39 (DMCA) is not clear about it. Hong Kong provides some exceptions for “reasonable use”. Yet one thing remains common throughout the world and that is the growing concern regarding copyright violation by various AI systems, be it “Getty Images v. Stability AI Ltd.”40, violations by “Open AI”, legal database use by “Ross AI”, music rights violations by “Anthropic AI” or the recent “Studio Ghibli” (Miyazaki) fever throughout the net world. Legal accountability of all the violations remains a grey zone due to lack of proper laws.
AI generated creation as “derivative work”
Berne Convention (1886) describes derivative works in Article 2(3) as “new creative work based on or derived from existing works” and copyrighting such works require distinct difference from the original creation and not merely rearrangement of previous data. The insight of AI to generate new creation is well analysed by Dr Shlomit Yanisky-Ravid and other researchers but the copyright of AI generated creation varies in different part of the world. In 2021 European Commission proposed first EU AI law. EU does not recognise AI as the author of creation. UK gives copyright to the person who arranges for the creation. Hong Kong does not see it in same frame as traditionally developed creations. In September 2021 “legal framework for AI” was passed by Brazil’s legislature. South Africa in fact was the first nation to give authorship to AI system. India’s the Digital Personal Data Protection Act, 2023 (DPDP Act, 2023)41 remains vague regarding AI generated creation as did the previous “Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021”42.
Conclusion
From “Statute of Anne”43 of 18th century Britain to “Convention of Berne”44 of 19th century Europe and from “WIPO Copyright Treaty (WCT)” of 20th century to 21st century novel issues of “BCI” based metaverse, the same question that puzzled the US Supreme Court in “Burrow-Giles Lithographic case”45 haunts us at the liminal dawn of a new age in a new form and makes us think over and again, whether an image perceived by the eyes of a person and later deciphered from the mind of that person using a machine be copyrighted in the same way an image caught using fingers of that person is copyrighted? This question becomes more important as we enter the field of “disability rights” where using machine is not a choice but a need. With advancement of science, now we know that our eyes are nothing but an advanced bio-camera and our brain a very advanced data processor and bio-databank. The techno-centricity of an anthropologically disabled person’s creation puts the archaic anthropogenic copyright laws in question and demands a discourse on this aspect of utilitarianism in an egalitarian world vision. Certainly, the vision of egalitarianism faces the hurdle of affordability with digital divide seeding a nidus of neo-colonial system where intellectual property right of technocentric creations get ensnared within the dilemma of acceptability and affordability. The disability of the Copyright Acts in addressing the intricacies of copyright issues inherent to disability rights of a necessity-based man-machine hybrid takes us to the transhumanist reality of posthumanism. The ethical basis of post-human laws and the IPR issues of man-machine interface does not lie in mind-controlled drones of warfare or mind-controlled video games, but it lies in the need-based development of such technologies for realising the disability rights of the present era. “A Brief History of Time”46 not only gave us a clue to the origin of the universe, but it also gave hope to a million disabled persons who saw a way to express, create and author using technological advancements like BCI based AI models and thus harbingered a new era of copyright dilemmas. As “neurotechnology” advances, the right step towards developing a conducive world surrounding it would be to balance anthropocentric rights with technocentric aspirations as stated in “León Declaration on European Neurotechnology”47 in October 2023 and thus it would be a step towards protecting IPR of BCI generated creations in an era of virtual reality (VR). Steps are also needed to ensure “privacy rights” and provision of “consent” described in “General Data Protection Regulation” (GDPR)48 of EU to include that of BCI generated data as well. Present copyright laws do not answer the novel questions that the BCI era asks and in this new era concepts on the intangible mind needs to be protected to secure the tangible creations of the BCI-AI interface using our mental hues.
Note: This article was selected as one of the winners in an essay writing competition organised by the Centre of Artificial Intelligence and Intellectual Property Rights, Himachal Pradesh National Law University, Shimla.
*National Law School of India University, Bangalore. Author can be reached at: ani950id@gmail.com.
1. A. Valjamae, “We Should Protect the Intangible Aspects of Our Mental Worlds”, European Science-Media Hub (sciencemediahub.eu, 31-1-2024).
2. J.J. Vidal, “Toward Direct Brain-Computer Communication”, (1973) 2 Annual Review of Biophysics and Bioengineering 157-180.
3. A.M. Turing , “Computing Machinery and Intelligence”, (1950) 59 Mind 433.
4. Yanisky-R.S. et al., “When Artificial Intelligence Systems Produce Inventions: The 3A Era and an Alternative Model for Patent Law”, (2018) 39 Cardozo Law Review 2215-2263.
5. 1884 SCC OnLine US SC 113 : 28 L Ed 349 : 111 US 53, 55 (1884).
6. 2018 SCC OnLine US CA 9C 94.
7. Copyright Act, 1976, § 101.
8. 1884 SCC OnLine US SC 113 : 28 L Ed 349 : 111 US 53 (1884).
9. Genesis 3:19, “By the Sweat of Your Brow You Will Eat Your Food Until You Return to the Ground, Since from it You Were Taken; for Dust You are and to Dust You Will Return”, The Bible Hub [New International Version (NIV)]. See also, judgment by “House of Lords” in Walter v. Lane, 1900 AC 539 and judgment in Cummins v. Bond, (1927) 1 Ch 167. See also, Cousins, Wendy, “Writer, Medium, Suffragette, Spy? The Unseen Adventures of Geraldine Cummins”, (2008) 45 The Paranormal Review 3-7. See also, Ladbroke (Football) Ltd. v. William Hill (Football) Ltd., (1964) 1 WLR 273.
11. Cummins case, (1927) 1 Ch 167.
12. Flavour of minimum requirement of creativity, Eastern Book Co. v. D.B. Modak, (2008) 1 SCC 1.
16. 1879 SCC OnLine US SC 4 : 25 L Ed 841 :101 US 99 (1879).
17. 1954 SCC OnLine US SC 23 : 98 L Ed 630 : 347 US 201 (1954).
18. “Scene that must be done” (French) cited in Cain v. Universal Pictures Co., Inc., 1942 SCC OnLine Dis Crt US 1. See also, J. Beeber and M. Wogan, “Is Scènes à Faire Really ‘Necessary’?”, 15(1) Entertainment, Arts and Sports Law Journal (Spring 2004).
19. Sweat of brow doctrine was rejected by US Supreme Court in Feist Publications Inc. v. Rural Telephone Service Co. Inc., 1991 SCC OnLine US SC 46 : 113 L Ed 2d 358 : 499 US 340 (1991).
20. Copyright Act, 1957, S. 13.
21. Copyright Act, 1976, 17 USC, §§ 101-122.
22. Berne Convention for the Protection of Literary and Artistic Works, 1886, Art. 3(1), 331 UNTS 217 (1886).
23. Copyright, Designs and Patents Act, 1988 (GB).
24. 1973 SCC OnLine US SC 135 : 37 L Ed 2d 163 : 412 US 546, 561 (1973).
25. Goldstein case, 1973 SCC OnLine US SC 135 : 37 L Ed 2d 163 : 412 US 546, 561 (1973).
26. Agreement on Trade-Related Aspects of Intellectual Property Rights (TRIPS), 1995, Art. 9(2).
27. WIPO Copyright Treaty (WCT), 1996, Art. 2.
28. 1938 Ch 106 : (1937) 3 All ER 503.
29. Donoghue case, 1938 Ch 106 : (1937) 3 All ER 503.
31. Barbara Taylor Bradford case, 2003 SCC OnLine Cal 323.
32. 1884 SCC OnLine US SC 113 : 28 L Ed 349 : 111 US 53, 55 (1884).
33. Copyright Act, 1957, S. 37.
34. 1841 SCC OnLine Dis Crt US 1.
35. Copyright Act, 1976, S. 107. (United States)
36. 1994 SCC OnLine US SC 22 : 127 L Ed 2d 500 : 510 US 569 (1994).
37. 2015 SCC OnLine US CA 9C 1.
38. Copyright Act, 1957, S. 52.
39. Digital Millennium Copyright Act, 1998 (US).
41. Digital Personal Data Protection Act, 2023.
42. Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021.
43. The Statute of Anne, enacted in Great Britain in 1710, is widely considered the first copyright statute.
44. Berne Convention for the Protection of Literary and Artistic Works, 1886, Art. 3(1), 331 UNTS 217 (1886).
45. Burrow-Giles Lithographic case, 1884 SCC OnLine US SC 113 : 28 L Ed 349 : 111 US 53, 55 (1884).
46. Stephen Hawking, A Brief History of Time (Bantam Books, 1988).
47. Spanish Presidency of the Council of the European Union, “León Declaration on European Neurotechnology: A Human Focused and Rights’-Oriented Approach”, (Leon, 24-10-2023).
48. General Data Protection Regulation (GDPR) (gdpr.eu).
