The recent order of the Delhi High Court in Stephen Thaler v. Union of India to expeditiously decide on an application filed by Stephen Thaler seeking copyright registration for ‘A Recent Entrance to Paradise,’ an artwork generated by his AI system, DABUS. While it does resolve the substantive question of whether artificial intelligence can generate copyrightable works, to treat it as a routine procedural direction would be to miss its larger significance. By directing the Copyright Office to decide a long-pending application within eight weeks, the Court has effectively foregrounded a pressing doctrinal gap in Indian copyright law.
Justice Tushar Rao Gedela’s order is deliberately restrained. The Court refrains from engaging with the merits of Thaler’s claim and confines itself to addressing administrative delay. Noting that the application had been pending since 2022 without meaningful progress, the Court emphasised that such inaction undermines the efficacy of statutory remedies. In doing so, it reaffirmed a familiar principle of administrative law, namely statutory authorities must act within a reasonable timeframe, particularly where emerging and unsettled legal questions are involved. However, the Court’s insistence on a time-bound decision carries a broader institutional signal. It recognises that questions of AI authorship are no longer speculative or academic, but they are immediate, practical, and demand regulatory clarity. By compelling the Copyright Office to decide the matter, the Court has ensured that the substantive issue can no longer be deferred indefinitely.
The Indian Statutory Framework: Room for Interpretation
The governing provision in the Indian statutory framework is Section 2(d)(vi) of the Copyright Act, 1957, which deals with ‘computer-generated works.’ It provides that the author of such works is ‘the person who causes the work to be created.’[1] On its face, this language appears adaptable to AI-generated outputs. Unlike statutes that explicitly require human authorship, the Indian framework leaves open the possibility of attributing authorship to a human actor involved in the creation process, even if the immediate output is generated by a machine. Thaler’s argument seeks to capitalise on this flexibility. He contends that as the developer and operator of DABUS, he satisfies the statutory requirement of being the person who ‘causes’ the work to be created.[2] The autonomy of the AI system, in this framing, does not negate human authorship, rather, attributes creativity to the design and deployment of the AI system itself.
This interpretation, if accepted, would significantly expand the scope of copyright protection in India. It would allow works generated with minimal or indirect human input to be protected, provided a human actor can be identified as the causal agent. Yet, this reading is not without difficulty. The provision was enacted when ‘computer-generated works’ referred to outputs generated through processes under human control, not systems capable of independent generative behaviour.
The Comparative Landscape: Convergence and Divergence
In the United States, courts and administrative authorities have adopted a firm and consistent position that copyright protection requires human authorship. Thaler’s attempt to register the same artwork in his AI’s name was rejected by the U.S. Copyright Office, a decision upheld by federal courts.[3] The Supreme Court of the United States ultimately declined to hear his appeal, effectively bringing the claim to an end.
By contrast, China, has adopted a more pragmatic approach. In a 2023 decision, the Beijing Internet Court, recognised copyright in an AI-generated image created using Stable Diffusion. However, it did not attribute authorship to the AI system itself. Instead the Court attributed authorship to the human user who input prompts, adjusted parameters, and curated the final output. The emphasis was on ‘intellectual achievement’ and the presence of human choices reflected in the final output.[4] In this respect, the Chinese approach can be distinguished from Thaler v. Perlmutter decision in the US, where Thaler sought recognition of the AI system as the author, rather than the human operator using AI as a tool.[5] This distinction is also important when comparing U.S. proceedings with those brought in India. In India, Thaler’s claim proceeds on the basis that he is the author using AI as his tool, whereas in the United States he advanced the more radical position that the AI system itself should be treated as the author. The latter position would also sit uneasily with the wording of India’s Copyright Act.
The United Kingdom presents a more nuanced statutory framework. Its copyright law also recognises computer-generated works and attributes authorship to the person making the necessary arrangements.[6] The term ‘computer-generated’ is defined as referring to works ‘generated by computer in circumstances such that there is no human author of the work.’[7] However, in practice, UK law has not departed from a human-centric approach. Courts have not recognised AI systems as authors, and the requirement of human involvement remains central.[8] At the supranational level, the European Union lacks specific rules on the copyrightability of AI-generated works, but existing case law of the Court of Justice of the European Union and developments at Member State level demonstrate a strong need for human creativity. The European Parliament has endorsed a human-centric approach while calling for further analysis of how intellectual property rights are applied to content created with the use of AI tools.[9]
This comparative landscape reveals a spectrum of approaches, involving different degrees of human input. At one end lies the U.S. model, which strictly excludes non-human authorship. At the other lies the English approach, which is willing, at least theoretically, to recognise copyright circumstances where there is no human author of the work. While there is statutory language that allows for a broad interpretation, it has not been followed in practice and yet to be fully tested in the context of autonomous AI systems. China occupies a middle ground, where copyright is granted where human involvement in AI-assisted creation meets the threshold of an ‘intellectual achievement.’[10]
The Delhi High Court’s order in Thaler v. Union of India may be limited to directing a time-bound decision, but its implications are far from narrow. By compelling the Copyright Office to act, the Court has set the stage for India’s first meaningful engagement with the question of AI-generated authorship. The resolution of these questions is not purely doctrinal, but also involves policy considerations. Expanding copyright protection to AI-generated works could incentivise innovation and investment in AI technologies. It could also provide legal certainty to industries increasingly reliant on generative tools. However, such expansion carries risks. Granting protection to works with minimal human input may dilute the concept of originality and lead to an over-proliferation of rights. It could also raise concerns about monopolisation of machine-generated content and the displacement of human creators. Whether the eventual decision aligns with the U.S. model of strict human authorship, adopts a more flexible interpretation akin to China, or charts a distinct path grounded in India’s statutory language remains to be seen. What is clear, however, is that the issue can no longer be deferred.
