The Dialogue’s latest study draws global lessons to propose a pragmatic, innovation-friendly framework for AI and copyright policy
New Delhi | December 26, 2025 – As India positions itself as a major global hub for artificial intelligence, a new report by leading public policy think tank The Dialogue has brought renewed focus to one of the most complex and consequential questions facing the country’s digital future: how to balance rapid AI innovation with the protection of copyright and creative rights.
The report, titled “Policy Prescriptions for Balancing AI and Copyright Concerns,” argues that India is at a critical inflection point. Decisions taken now, it warns, will determine whether the country emerges as a competitive AI powerhouse or becomes constrained by rigid regulatory frameworks that may inadvertently slow innovation while failing to meaningfully protect creators.
A Global Lens on a Local Challenge
At the heart of the report is an extensive comparative analysis of how major global jurisdictions are grappling with the intersection of AI training and copyright law. The Dialogue examined legal and policy frameworks in the United States, European Union, United Kingdom, Canada, China, Japan, and Singapore, offering Indian policymakers a panoramic view of how different countries are navigating similar challenges.
The study highlights that there is no single, universally accepted model for regulating AI training data. Instead, countries have adopted a spectrum of approaches—ranging from broad fair use doctrines and text-and-data-mining (TDM) exceptions to more restrictive, rights-holder-centric regimes. Each model, the report notes, brings its own set of trade-offs.
For instance, the United States relies heavily on judicial interpretation of fair use, which has historically enabled innovation but also created uncertainty for both AI developers and rightsholders. The European Union, on the other hand, has opted for explicit TDM exceptions with opt-out provisions, providing more clarity but also introducing compliance complexities. Japan and Singapore have taken comparatively innovation-forward approaches by offering broad statutory exceptions, while China has leaned toward tighter regulatory oversight aligned with state priorities.
By mapping these global approaches, the report seeks to ensure that India learns from international experience without importing regulatory frameworks that may be ill-suited to its economic scale, creative diversity, and developmental priorities.
Policy Prescriptions Tailored for India
Building on its global review, The Dialogue lays out a set of forward-looking policy recommendations designed specifically for India’s fast-evolving AI and creative ecosystems.
A central recommendation is the introduction of a broad, technologically neutral statutory exception for text and data mining, covering both commercial and non-commercial AI training. According to the report, such an exception would provide much-needed legal certainty for Indian AI developers while aligning India with other innovation-oriented jurisdictions.
The report also suggests exploring voluntary opt-out mechanisms, allowing rightsholders to signal preferences without imposing blanket restrictions that could fragment training datasets. In addition, it emphasizes the importance of metadata standards and attribution frameworks to promote responsible data sourcing and greater transparency in AI development.
Rather than advocating heavy-handed enforcement, The Dialogue calls for a guidance-first regulatory approach, prioritizing best-effort compliance and industry collaboration over punitive measures. The report argues that overly prescriptive or retroactive regulations could deter investment, especially for startups and smaller firms that form the backbone of India’s innovation ecosystem.
Addressing Real-World Harms at the Output Level
Importantly, the report draws a clear distinction between upstream training processes and downstream harms. It urges policymakers to focus regulatory attention on tangible risks such as deepfakes, voice cloning, misinformation, and unauthorized use of likeness, rather than attempting to micro-manage how large AI models are trained.
According to the authors, output-focused regulation is more effective, enforceable, and proportionate. By targeting demonstrable harms, the law can protect individuals and creators without placing unrealistic burdens on AI developers.
Practical Guidance for Industry and Creators
Beyond policy recommendations, the report also outlines actionable steps for both AI developers and rightsholders. Developers are encouraged to strengthen dataset provenance practices, proactively filter out clearly pirated content, and pursue selective licensing in high-risk areas. These measures, the report suggests, can significantly reduce legal exposure while building trust with the creative community.
For creators and rightsholders, the report advises a shift in enforcement strategy—away from broad claims against training datasets and toward evidence-based action against harmful or infringing outputs. It also highlights the potential of collaborative and co-creation models, where AI companies and creators work together to unlock new revenue streams and creative possibilities.
Industry Voices Weigh In
The release of the report sparked a robust discussion among industry leaders, particularly around the recommendations of the recently released DPIIT report, which proposed a hybrid mandatory licensing framework supported by a government-led rate-setting body.
Suparna Singh, CEO of Frammer AI, cautioned against approaches that could undermine India’s global competitiveness. “Enabling innovation and maximum benefit for all must be the guiding principles for India’s AI–copyright framework,” she said. “Mandatory licensing or government-determined pricing, however well-intentioned, rarely work. The market should determine what data is valuable and how much it should cost.”
Echoing similar concerns, Sanjay Sidhwani, CEO of The Indian Express Digital, highlighted the complexities of valuing content, particularly in the news industry. “There is broad alignment on acknowledging copyright and the need for compensation, but a one-size-fits-all framework cannot reflect reality,” he said. “The value of a century-old archive is vastly different from recently created content, yet proposed models struggle to make such distinctions.”
Venkatesh Krishnamoorthy, Country Manager-India at the Business Software Alliance (BSA), emphasized the link between AI adoption and innovation. “India’s AI strategy must focus on accelerating adoption across sectors of all sizes,” he said. “Without adoption, innovation stalls. At the same time, assumptions that individual works used in large-scale model training can be precisely identified and retrospectively valued raise serious feasibility concerns.”
A Call for Multistakeholder Dialogue
A common theme across the panel discussion was the need for sustained multistakeholder engagement before finalizing any policy direction. Industry leaders, creators, policymakers, and civil society, they agreed, must work together to design a framework that reflects technical realities, economic incentives, and public interest.
Through this report, The Dialogue aims to move the AI-copyright debate away from zero-sum narratives and toward constructive, evidence-based solutions. As India stands on the cusp of an AI-driven transformation, the study underscores that the choices made today will shape not only the future of innovation, but also the country’s creative economy and democratic digital ecosystem.
The full report, “Policy Prescriptions for Balancing AI and Copyright Concerns,” is now available on The Dialogue’s website.
