Balancing Creativity and Control: Analyzing Intellectual Property Laws and Global Approaches to Regulating Artificial Intelligence

Introduction

Artificial Intelligence (‘AI’) came as a transformative technical revolution, which like any other previous revolutions in human history made significant change in human lives, but it has also illuminated new dangers in the socio economic and ethical lives of the peoples who holds the traditional intellectual property rights.  While there is no certain and definite history of ‘AI’, John McCarthy in 1958 created LISP (acronym for List Processing), the first programming language for AI research and coined the term ‘artificial intelligence’ in a workshop hosted at Dartmouth. Release of open plat forms like ChatGPT in late 2022 has put this technology back under the spotlight.  With these   rapid growth in the field of ‘AI’ particularly in the domain of Generative Artificial Intelligence (GAI) , continues to redefine the boundaries of technological capability and future development of technology. These technical developments made AI capable of generating both audio and visual content from use of simple text prompts, these innovation looks like the life of any science fiction movie which have now incorporated in day to day lives. These modern advancements posed challenges to the longstanding assumption that ‘only Human can create and innovate’.[1]AI is completely in its nascent phase and is continuously evolving, much of its legal and ethical Empact on the copyright regime and its impact on right holders remains to be seen. While AI not only always poses harms on the creators but also provide creative assist to the creator. In this article a detail examination between the utility and threat of AI – generated art is presented. This article will also investigate the potential revenue disrupt of creative industries, from the collection of dividends from intellectual property rights. This article also explores the laws and regulation which have been incorporated in the Indian legislations to deal with the copyright matters and other ownership rights. Subsequently this paper reviews the legal questions raised by different shareholder related to the real threat proposed by the AI generated outputs in the area of intellectual property. Finally, it proposes a new IP framework based on officially “licensed AIs”, viewed as a new format where the IP holders can collect AI royalties. The most important question that revolves around is whether AI should be regarded as author under copyright regime, answer of this question is covered under 2 different set of answers, one that favors it and other present optimistic views. This paper will work as a guiding filed and provide critical analysis for different shareholder like law students, scholars, policymaker.

Legal Framework of IPR in INDIA –

Indian Law recognizes several classes of IPR that protect creative expression. These includes:

Trademark: The Trademark 1999Act protects the brand names, logos, and other identifiers of source. Additionally, India is also a signatory to TRIPs agreement which protects trademark from its article 15 to 21.

Copyright: Copyright is a unique kind of intellectual property. The right which a person acquires in a work, which is the result of his intellectual labour. Copyright is protected under Copyright Act ,1957 where the literary, dramatic or musical works; artistic work; cinematograph film; and sound recording are protected.

Neighboring Rights: According to WIPO neighboring rights, protects the other rights of performs (actors/musicians); Producers of sound recordings (also referred to as phonograms); and Broadcasting organizations which are not explicitly covered under copyright act.

AI as Author- These traditional laws provide a wide and comprehensive set of rules for protecting the traditional form of intellectual property but this new era raises a new question of whether AI can be considered as Author of an intellectual work? Authorship has evolved as different concept across the various jurisdiction.[2] majorly due to the differing standard of originality in finding the copyright of the work. Various approaches have been adopted by various courts to find creativity in the works. Despite the differences in the way of finding authorship across various jurisdiction one thing remained always same that is the authorship is exclusively reserved for the human species.[3]

3. Outlining the concept of AI with related IP laws-

3.1 The term AI is unfathomable, different authors defined it differently. For stance American technology company IBM states AI as “any system capable of simulating human intelligence and thought process is said to have ‘artificial intelligence’”.[4] AI is based on the study of algorithms of computer that automate analytical model building and it is referred as Machine learning. AI is like a software which is trained by the humans using different language modules (LLM’s) after which when the prompt is given to the system it generates the material by itself from the pre available data and provide it as output. These outputs are totally un predictable which causes the serious concern whether the creation by AI is eligible for the copyright.

3.2 Creator vs owners-

In simple wordings a creator in the first and the initial owner of the copyright, which grants them the exclusive rights to their work. How ever the creator can transfer their rights by the agreement, to whom the rights is transferred becomes the owner of the rights however the creator do not lose certain rights even after transfer such as moral rights and right to identified as the author. The question of the creator vs owner is raised here for two different purposes the first is, whether the system that is creating art (company or corporation that trains the AI) should be known as the author of the art or the person who is giving the command to AI should be the author of the art? The second question is whether Data used for training the AI system is infringement of the copyright or not. The New York Times Co. v. Microsoft lawsuit highlights one such problem and decision by the court. In this case the New York Times alleged that open AI created by Microsoft uses the data of ‘times’ prewritten copyrighted material published online without permission to train its AI. The Times not only concerned with the ownership right it secured but also with maintaining a greater share of profits it reaped from those initial contracts.

4. Statement of problems

While dwelling with the common problems this paper revolves around some common questions raised on internet and other surveys –

Whether AI should be considered as Author under copyright regime and what will be the copyrightability of the work generated by AI?

Whether AI is a threat to the traditional artists or it works as positive development in the field of art?

Whether there are sufficient laws available to regulate the process of training AI is so that intellectual property rights can be prevented from infringement?

Whether providing creative freedom to use AI in different art be a faceable option?

5. Whether AI should be considered as Author under copyright regime and what will be the copyrightability of the work generated by AI?

5.1 With the new era of artificial intelligence this new question also comes into picture where different intellectuals take different stands. First view talks about status quo where Humans remain the sole creators and say that since the AI is creation of human intelligence, the output generated by AI is simply a ‘copy paste’ of pre-existing materials created by humans so AI can not be considered as the author of any art work and cannot hold right over intellectual property. Tis view is further supported by the jurisprudential principle of ‘juristic personality’, since AI can not hold the duties towards other it can not be held eligible for the rights. On the other hand

Some supports the justification for recognizing AI as author these justifications are broadly given on two principles, first one is right based, which is based on John Locke’s Labour theory as well as the personality theory propounded by Hegel. Other one is of utilitarian based of the concept of J. Bentham. The former gives its justification that AI uses sufficient labour to performs its task and the output is the pure result of its labour therefore the AI is the owner while later says that the work is an expression of the author’s own personality-an embodiment of its will.

5.2 Another group of people justify this on the ground of emerging consciousness in AI systems which is embedded in the worth and sacredness of non-human life and forms. Research from various disciplines has added towards the creation of a global ethic for safeguarding and fostering the non-human realm. It seems your input was cut off. the statement by Svoboda, “non-human entities possess intrinsic value independently of their value derived from usefulness. Svoboda states that "intrinsic value independent of the mind is considered to be a property held by certain entities regardless of the beliefs, desires, or attitudes of anyone genuine or potential evaluator or cognizant individual.” [5]

If we consider AI as a juristic person capable of holding rights then we find that AI fulfills all the traditional elements of an author required under laws of copyright in India, the output provided by the AI have substantial amount of creativity and also provides original work hence it fulfills all the grounds of statutory requirements.

5.3 Benefit of granting copyright to AI-generated or AI assisted works-

There are many problems with grant of copyright or any other intellectual property right to AI generated or AI assisted works but here are few benefits also, one of these benefit is if AI works will hold the right over its intellect then the creator of the AI or the owner of the AI will have to show the language or other modules (like LLMs) used to train AI and this will create the transparency between the AI and its user which will further prevent the unauthorized use of the copyrighted work and prevent infringement.

6. Whether AI is a threat to the traditional artists or it works as positive development in the field of art?

This is further a question in the series that is answered differently by different groups according to their interest or their expertise. In a survey conducted by University of Vermont Institutional Review Board (IRB)[6] where 459 artists were surveyed out of them 69.87 percent agreed that AI models are threat to art work only 44.48 percent of them agreed that AI art models are a positive development in the field of art. This survey though represents only a small data but shows that artist who believes that AI is a threat to art are more that who do not believe, one of the most important reasons for this issue is opaqueness in training the AI models. While the other group argues that since human artist have adopted many scientific tools from time to time, AI is not separate from other. AI is a just a software with its own limitation that are prompted by the AI companies so it works under that circle only and can not be a threat to the real artist. It can enhance the productivity of the artist by assisting them on given prompts, US district court of Columbia also agreed from this in a case of 2023 where court denied the copyright of AI generated art underscoring role of human creativity in copyright laws. 

6.1 one recent example of use of Ghibli art by ChatGPT (AI), Ghibli art is a traditional painting originated in Japan by Hayao Miyazaki, Studio Ghibli owns all the rights to use of Ghibli style painting but recent AI ChatGPT came with a feature of creating similar art without license. Which clearly shows how AI platforms can be misused specially when the training of platform is not disclosed. Use of AI without giving credit to the artist is not only creates infringement of moral rights but also the economic rights of the artist for their copy righted work.

6.2 AI models can be both positive developments or threat to artist it depends upon the use of AI and the laws to regulate the AI, there should be human centric relation between the machines and humans, ‘the machines should not use humans, only human should use machines.

7. Whether there are sufficient laws available to regulate the process of training AI is so that intellectual property rights can be prevented from infringement?

Though AI is not a new concept for the World including India, probably the potential of AI is realized by the Indian legislation framers recently when the real threat of AI is realized in different fields including Intellectual property Rights.

7.1 Currently Indian legislations have legal Vacuum; Indian copyright office have created many temporary policies to deal with such issues but they do not fully comply with the need of today.

Indian courts have tried to fill this legal lacuna through judicial pronouncements. Overall Indian legislation has yet to issue such definitive guidelines, to deal with issues with are created by the rapid development of AI.

7.2 Lacuna in Indian Legislations, Absence of legal clarity-

The Copyright Act 1957 or Trade Mark act 1999 does not define AI-generated works though it provides inclusive term of “computer generated work” but this is ill-equipped to cover modern AI-Generative models that works on autogenerating and auto learning capabilities.

7.2 Work by courts- Indian judiciary have tried to collaborate or fulfill the vacuum but rapid growth of AI in the field of journalism, digital art, entertainment, etc. have over powered the current legal system and shows how Courts have not taken up issue of AI authorship in any substantial way.

7.3 Comparative checks- Though the area of AI authorship remains largely untouched by most of the legislations of the world but some of these act tries to cover this to some extent, The UK Copyright, Designs and Patents Act 1988 provides a unique provision in Sec 9 (3) for computer-generated works, the author is "the person by whom the arrangements necessary for the creation of the work are undertaken." This pragmatic approach addresses the issue of AI authorship, although it too assumes a human in the loop.

United State’s Copyright office and court have rejected the protection to the AI generated works affirming to the traditional approach of granting intellectual property licenses to the holders. The decision of the U.S supreme court in the case of Bridgeman Art Libr., Ltd. v. Corel Corp followed the same directives of US courts but also open a door for consideration. Court stablished the principle that level of originality will be depend on the context in which the work is produced, for example the photograph of toys was treated as “original works of authorship” though provided by AI but digital copies of physical art work were not granted the same treatment. With this decision of court, we may assume that U.S supports the art generated by AI for copyright were there is minimal involvement of AI and most of the work is the result of human creativity and labour of Human, this shows how US have adopted the developments in AI with respecting traditional rights of intellectual property.

8.0 Whether providing creative freedom to use AI in different art be a faceable option?

History shows us how new technologies have from printing press to camera have always helped the artiest to adapt and develop new creative possibilities, AI may be the another one. For some creators like writers and musicians, AI opens many options to develop and expand their capabilities but it also poses new threat to their concept of human creativity as AI now can develop inherently new scripts, songs, etc.  AI creates thousands of possibilities of generating content but no matter how complex or creative future AI models become, they will never be able to literally generate a human to perform their work. Many studies shows that AI can not perform over what human allows them to perform no matter how developed AI is it will always be subject to the human bonds. So, allowing the use of AI in different art will not hamper the copyright of human but it will enhance their creative freedom if used in a legally bound and regulated manner.

9.0 Recommendations and the Way Forward

9.1 Granting Copyright protection to AI-assisted works and Preventing copyright to AI generated works- AI generated work can be referred as those works where user has made only de minimis contribution while giving the input that is the output generated by AI is sole outcome of AI intelligence and thinking ,in this case the authorship should be granted to AI not the user and this should not be the subject matter of copyright. On other hand AI assisted works are those works where AI has done only minimal task and the work is the result of labour and creativity of the user. In a similar case of U.S. Copyright Review Board, Mr. Sahni used a generative AI tool called RAGHAV to do some modification in original photo graph, to transform his photograph’s artistic expression by giving instructions to AI to apply a different style in photo. The inputs he provided to the AI included the original image, a reference artwork (notably The Starry Night), and a numerical value to guide the intensity of the style transfer. The Board found that Mr. Sahni did not contribute any expressive element in the final art. Hence can not be the sole author of the art. In this case the Art r the final out put was the sole creation of AI and sole work of AI hence Mr. Sahn is not the author of the art. since the creative decisions—such as how the style was applied and the images were combined—were made entirely by the AI based on the provided parameters, authorship could not be attributed to him.

9.2 Need for Legislative Amendments

With the rapid change in the technology and with this development in AI legislations should also change to cope up with new threat and problems. Indian legislation has now yet opted any regulation to specifically deal with this issue, now the time has come when India must amend the Copyright act and other related intellectual property act to adopt clear regulation, for example defining ‘AI-generated and ‘AI-assisted’ work in act and determining the new criteria of authorship for these works.

Assigning clear intellectual property rights to the user and developer of AI so that legal creativity could be maintained.

Creation a sui-generis legislation for AI generated content or inclusion of this preexisting act could be a possible way forward for new world problems of AI. Help of experts can be taken for inclusion of this type of legislative provisions so that interest of all can be maintained.

Most important step that should be taken by the government is to introduce schemes for training data usage, also create the transparency in the data used by AI developers to train AI this will ensure the unauthorized use of copyrighted data on online platform by Ais for their training or generating outputs, in case of use insuring the moral and economic rights of original creator.

9.3 Special protection for Human creator

The governing authorities by issue guidelines can make the provision for disclosure label that should be carried out by the works produced by AI to prevent infringement and also stablish sufficient amount of transparency. This will specially ensure profitability in the field of mass communication where the AI is used directly to generate content.

Creation of an IP framework to compensate original artist or other IP holders for AI generated work- there should be a legal and transparent compensation model that treats “licensed AIs” as a form of creative property and enables artist and right holders to earn royalties from AI system on their authorized works and this system will also analyze the economic and technical viability of IP holders and artists. One such model is created by Massachusetts Institute of Technology, US. [7]

With the advent of new generative artificial intelligence (AI) models– specially focusing generative adversarial networks (GANs) (2020), VQVAE (2021) and probabilistic diffusion models (DMs) (2015, 2022), production of AI generated content became too fast ushering new area of high quality of multimedia which is generated by AI it becomes important to create new form of model which can regulate the royalties for AI generated work.

9.4 Role of Judicial interpretation –

Judiciary can paly a crucial role in maintain the balance between traditional and new form of IPs. Judiciary will serve the best of human interest in absence of legislative clarity on AI – generated content. We have seen the judiciary using doctrine of ‘purposive interpretation’ where there is involvement of technology and fundamental rights and no specific regulations are there to deal with such issues. Noably in case of 2017 Justice K.S. Puttaswamy v. Union of India Supreme Court adopted the doctrine of purposive interpretation and recognized right to privacy as fundamental right when there was no explicit mention of any such right in constitution. Judiciary can also ensure judicial sensitivity to new technological evolutions so that misuse of these evolution can be prevented.

Similarly, in the case of AI-generated works, courts may be required to adopt a dynamic and purposive interpretation of the Copyright Act, 1957, particularly sections dealing with “authorship” (Section 2(d)), “originality,” and “infringement” (Sections 51–63).

9.5 One crucial way forward for this issue can be forming the groups and working as a cohesive unit and take their decision which benefit for all like writer’s union, singer’s union. This can work as pressure groups and can raise there demands directly to the authorities. This step would provide them a way to effectually negotiate contract provisions that both allowed AI in writer’s room and guaranteed that AI would not be used in the way that can dilute their original creations or profits. Further they can stipulate that their copyrighted work would not be used without their permission for training LLMs and if work is used with permission they can further negotiate for the compensation. Hence working in groups can help the creators in two folds firstly for protecting their work (protection) secondly, for a piece of the revenue stream.[8]

 9.6 Private law can act as the secondary protection system for the IP right holder’s community by forming contract, licensing, trust, etc. this can insure the use of their work in LLMs and other potentials.

10 Conclusion

Human had a great history of adopting new technologies and forming new technologies for the purpose of convenience, but history also tells us that if these technologies not used in right manner, then it has created a reversed impact on humans. It is said that “history repeats itself, if not in same way, then in a similar way”. Similarly, this new era of artificial intelligence has created a global buzz in a revolutionary way where it has impowered humans to work with great efficiency and minimal mistakes, but if soon new regulations will not be adopted to regulate AI this could lead to a great disaster in many fields, one of the largest wills be Intellectual Property. In conclusion, there is no reason at all to devalue or undermine the case for recognizing AI authorship (and joint authorship) in copyright law, which this article has fervently sought to establish. Now, if only human authors are prepared to accept present day realities and warmly embrace their machine counterparts as equals. As far as the authors are concerned, the clarion call has been sounded. It is now up to Parliament to make the next move sooner rather than later.