Artificial intelligence—which Donald Trump has rebranded as “superintelligence”—is advancing at a rapid pace; hardly a day goes by without at least half a dozen news stories or opinion pieces on various aspects of the technology appearing in print media. It would come as no surprise to see even more unexpected developments emerge in the future.
It is little wonder, then, that the discourse in India surrounding AI governance is shifting from grand ambitions to difficult choices. The core issue is not merely how India should regulate AI, but rather what kind of AI economy such regulation will ultimately shape. This question has been brought into sharper focus by a recent Delhi High Court ruling in the case of ANI v. OpenAI. In July of this year, the court made a preliminary finding that OpenAI’s use of copyrighted news content from ANI to train its large language models (LLMs) might fall under the “fair dealing” exception provided for in Indian copyright law. In reaching this conclusion, the court looked beyond the narrow confines of copyright, weighing broader considerations such as innovation, access to information, the dissemination of knowledge, and the public interest.
The ruling has since sparked considerable debate among stakeholders, placing India among the growing number of jurisdictions currently grappling with how to regulate the training of large language models and the legal implications that follow.
However, as India’s artificial intelligence architecture begins to take shape, its core policy objective should be to determine the specific economic and innovation outcomes it aims to prioritize.
India is already investing heavily in building its domestic AI capabilities. The “IndiaAI Mission” is focused on reducing computing costs, expanding access to data, providing financial support to local startups, and improving the nation’s research infrastructure. Earlier this year, the Indian government established the AI Governance and Economy Group—a high-level, interdepartmental mechanism designed to coordinate the country’s AI policies, governance, and economic strategies. This ambitious vision now requires a corresponding legal framework to reconcile the various, potentially conflicting, regulatory goals that shape this ecosystem.
This is unlikely to be achieved through a single regulation; rather, it requires the coordinated operation of multiple legal domains. Copyright law will address the acquisition of protected training materials. Competition law will influence market access and market dominance. Data protection law will restrict the use of personal data in training datasets, while liability rules will determine the risks developers must bear when their models produce harmful outputs. Each legal domain targets different issues, yet their impacts ultimately converge on the same stakeholders and the same market.
The tensions surrounding copyright and artificial intelligence (AI) training highlight a complex intersection of interests. Authors, journalists, publishers, and other creators require viable economic incentives to continue contributing to the knowledge economy. Allowing AI developers to train commercial models on copyrighted material without the creators’ permission or compensation could undermine the economic vitality of the creative industries that produce such content.
Conversely, a rigid licensing regime carries its own economic costs and challenges. Training sophisticated models requires access to vast datasets. Licensing on such a scale entails numerous complexities—identifying copyright holders, negotiating terms with a multitude of creators and publishers, managing royalty payments, and monitoring compliance at every level. Global AI firms, backed by substantial capital and established licensing networks, might gain an advantage that local startups would struggle to overcome.
This creates a potential regulatory paradox: strengthening protections for creators could raise barriers to entry for local firms, thereby reducing competition in the AI market and entrenching the position of dominant AI companies.
Viewed in this light, the policy choice between permitting broad use of copyrighted material and mandating licenses for AI training has implications that extend well beyond the scope of copyright law. Consequently, as India evaluates its legal and regulatory framework for AI, it should consider not only how to address immediate legal concerns but also how its choices might shape market structures and economic incentives. Furthermore, India must incorporate sufficient flexibility to address unforeseen challenges that may arise as AI technology evolves.
The United Kingdom offers a useful precedent in this regard. In March 2026, the UK government published a formal “Copyright and AI Impact Assessment,” which evaluated how various copyright frameworks for AI training would affect copyright holders, AI developers, small businesses, and the broader economy. Crucially, the government linked potential copyright reforms to the extent to which a given approach could advance its wider economic and civic objectives.
The significance lies not in the UK’s final conclusion, but in its commitment to clearly defining desired outcomes before settling on a specific policy framework.
While India need not replicate this model exactly, it can certainly draw valuable lessons from the approach taken. First, India should clearly define the economic objectives it seeks to achieve through its artificial intelligence framework—such as sustaining creator incentives, preserving market competition, strengthening domestic AI capabilities, and advancing the broader public interest. Subsequently, before determining the specific legal provisions to be enacted, India should evaluate how various legal approaches might influence these objectives.
This outcome-oriented approach should extend beyond the scope of copyright. Interventions related to competition, data protection, and safety obligations must also be assessed for their cumulative impact on market access, investment, innovation, and public welfare. While the mandates of different regulatory bodies may vary, all AI-related rules should be evaluated against a unified set of economic goals to ensure that their combined effect supports—rather than undermines—India’s broader objectives for AI development.
The “AI Governance and Economy” group offers a practical institutional framework for India to undertake this task. It could establish a comprehensive economic impact framework—serving as a mandatory checkpoint for significant AI-related proposals—to test their potential effects on creator incentives, competition, domestic capabilities, and public welfare before such proposals are adopted.
The case against OpenAI signals a broader trend: existing laws will increasingly be called upon to address challenges arising from technologies that fall well outside the scope of their original design. While judges will continue to adjudicate individual disputes and regulators will craft rules within their respective jurisdictions, India bears a greater responsibility: ensuring that these decisions align with a coherent economic strategy for AI governance. The shape of India’s AI-driven economy should be the result of deliberate, well-considered choices, rather than something discovered only after the regulatory framework has already been set in stone.