Shortly after our first co-authored post on the copyright-related aspects of artificial intelligence and robot-journalism in India was published on this blog, we became aware of incredible news: the Indian Copyright Office has recognised the RAGHAV Artificial Intelligence Painting App and its legal owner as co-authors of the artistic works generated by the application. This blog post aims to give a short overview of the concerns related to AI-authorship (from a European perspective) and then summarizes the available information on this ground-breaking novelty of the Indian copyright system.

On AI, authorship and copyright law (Péter Mezei)

Source: Towards Data Science

The copyright protection of AI-generated outputs is nothing new under the sun. Policy debates, academic discussion and a slight, but existing case law has focused on the protectability of contents (we’d rather skip referring to them as “works”) that are generated either with the help of or by algorithms for several years and decades. The stakes are, however, high. It is not only about the doctrinal integrity of copyright law (what and who should protected; who rights, term and limitations should apply; what procedural role should be granted to algorithms etc.). Indeed, and even more, this debate seems to be about the “AI race”: who takes the first step in the copyright protection of AI-generated outputs to the benefit of new (and hopefully not for the detriment of earlier – in fact: human) creators. In 2020, a Chinese court ruling granted protection for AI-assisted news reports – although that protection was granted to Tencent, the publisher of the contents. (Another notable issue is the registration of DABUS as the co-inventor by the Patent Office of South Africa; but this is not a copyright issue.)

One of the most important recurring questions is related to the subject of protection: who should be granted protection for the creation of the output. (For the sake of clarity of this blog post, put aside my preferred default: NOONE; as I believe that such outputs shall belong to the public domain.) The options might range from the algorithm itself through the developer, the owner or the user of the algorithm to any kind of mixture of these possibilities. One of the key doctrinal concerns related to the protection of algorithms was the lack of personality; which also necessitates the refusal of originality, authorship and moral rights on the side of AI. At the same time, these concepts tend to be fictions. Rights are what we grant to rights holders who we declare to be eligible for protection. Copyright’s history is basically anthropocentric – but more and more exceptions are known, including deemed authorship (under the U.S. “work-made-for-hire” concept) or the related rights (typically in Europe) for the benefit of producers of sound recordings or motion pictures. (At the same time, forget about the protectability of the monkey selfie!)

Hence, the question was only when and in which country will the first ever copyright protection be granted to an algorithm. And the winner is…

The RAGHAV Artificial Intelligence Painting App and the AI authorship in India (Anushka Tanwar)

India becomes the first country to recognize an AI painting app – RAGHAV – as one of the authors of the painting titled “Suryast”. In order to create the painting, RAGHAV used Vincent van Gogh’s painting “Starry Night” and a photograph taken by Mr. Ankit Sahni, the owner of the AI-based app. This recent recognition of an AI painting app as one of the authors comes as a surprise as it has never been done before. Nevertheless, at the moment, the Indian Patents Act, 1970 and the Copyright Act, 1957 do not have any provisions to promote inventorship, authorship and ownership of an AI. The provisions of the Copyright Act are ambiguous regarding the authorship of a work that has been created by an AI without any involvement of humans. There exists a set of limitations on the existing provisions and legislation. The Copyright Office rejected the previous application for registration of the artwork where RAGHAV was named as the only author. However, in November 2020, the second registration application was accepted where RAGHAV was listed as a co-author along with Mr. Sahni, who submitted the copyright registration as part of a classified submission to the Indian Parliamentary Standing Committee on “protecting AI-created work as well as AI itself.” At present, it seems AI can only be recognized as a co-author along with its owner rather than the sole author of the work.

On 23 July 2021, the Parliamentary Standing Committee in its report made recommendations to review and amend the existing Patents Act and the Copyright Act to incorporate AI and AI-generated inventions and works. The Committee recognized the importance and relevance of AI and machine learning, especially during the current COVID-19 pandemic. These days, digital applications are playing an important role in responding to the pandemic by either helping everyone to communicate with each other, or by helping to keep a track of the number of people who have been infected or died due to the pandemic. Moreover, it stated that AI is highly beneficial for India’s economy and further facilitates the expansion of technological innovations. Keeping in mind all the advantages, the Committee recommended a whole separate category of rights for AI and AI related inventions in order to protect them under intellectual property rights.

I do not agree with the committee’s reasoning that AI and AI-generated inventions should be granted protection under the existing laws because “it would incentivize innovation and R&D thereby significantly contributing to creativity and economic growth of the country.” I believe even if AI-generated innovations are not given ownership/authorship, they will nevertheless contribute to the country’s economy. Therefore, I am of the opinion that there is no need for any sudden changes to the existing legislation regarding AI ownership/authorship. Furthermore, if the work created by an AI is literary, dramatic, musical or artistic, it must be original for it to claim ownership/authorship over a copyright. However, all the work generated by AI is dependent on its programming, restricting it to go beyond it. The AI is limited to make use of the information and resources it already has, making it incapable of generating anything original; its content is based on someone else’s work. For instance, in this case, RAGHAV used the existing painting of Vincent van Gogh and a photograph taken by Mr. Sahni. Even for derivative works, it is required for the author to show judgement and skill, which AI lacks. Moreover, in case of infringement, Section 51 of the Copyright Act, 1957 states that only a “person” can infringe a copyright. AI holds no legal status, therefore, an important question arises, in case of infringement by an AI or its work, who is to be held liable?

Nevertheless, keeping aside the debate of authorship of AI, the recent developments and initiatives made by the Indian government to promote AI in the country are commendable. They illustrate a strong and progressive future AI holds in the country.

~~~

Anushka Tanwar is a law student at University School of Law and Legal Studies in New Delhi. Ms. Tanwar is currently pursuing her internship at AIRecht, Amsterdam.  She has published in Transnational Dispute Management, Asia law portal, Kluwer Patent Blog, Kluwer Copyright Blog and has upcoming publications in the European Intellectual Property Review and the ERA-Forum. Her interests lie in Copyright law, Patent law and Artificial Intelligence and law.