AI copyright and law concept with magnifying glass and wooden blocks. (Image for representation)
Credit: iStock Photo
The public domain is the backbone of copyright law. The law functions on a delicate balance between the rights of creators, owners and users. It ensures that future creators can rely on previous works, once their protection period has expired, to generate new creations. Though the duration of copyright protection is considered long by many experts (life of the author plus 70 years in the USA, and life of the author plus 60 years in India), it is tailored to balance the market interests of both owners and creators.
Beyond the said period, a work is free for any user to use, adapt or remix as part of their creative work (provided it does not mutilate the previous work in any form or erode the moral rights of the creator). In 2026, the first set of Nancy Drew books and Disney’s 1930 animated film Mickey Mouse from The Chain Gang will enter the public domain. These works will then be available for any user to publish or adapt. In India, the works of famed singer G N Balasubramaniam will enter the public domain this year.
Public Domain Day is celebrated across cultures as a way to nurture creativity. The 2026 edition is unlike any other in the past. Over the last year, generative AI has expanded beyond the imagination of both its users and its creators. It has challenged existing norms of copyright law in ways no previous technology ever has. This may diminish the importance of the public domain in the future.
Generative AI, like its human counterpart, learns from existing material to create new works. This is not a problem per se, since copyright law was designed to promote learning and creativity after all. In its present form, the law does not distinguish between human learners and generative
AI platforms. The only difference is in the massive speed at which AI systems can train and create. At present, these platforms learn from both copyrighted and non-copyrighted material to generate further.
For instance, the previous year saw the famous Ghibli trend surface on the Internet, where social media users posted their own images styled after Studio Ghibli works. It would not have been possible for AI platforms to generate such images unless they had been trained on those works. The resulting creations could potentially challenge the market for the original works on which they were trained. This sparked a copyright controversy, though nothing emerged of it. At present, copyright-respecting countries are responding to such ‘AI training’ through their respective policies.
In this process, India has signalled a stand that would seemingly balance the interests of generative AI platforms and copyright owners. It aims to allow AI platforms to train on both copyrighted and non-copyrighted works by obtaining a blanket licence from a body that would manage copyrighted material in exchange for a statutorily fixed royalty fee. This complex response as suggested by the Department for Promotion of Industry and Internal Trade in a recently released white paper.
Somehow, India’s response challenges the existing idea of the ‘public domain’. This has the scope to alter the future of copyright law. At present, users can adapt or reuse works only after the expiry of copyright; however, the proposed Indian measure would change this. It suggests that generative AI platforms could train on works even before their entry into the public domain. Copyright owners or creators, meanwhile, cannot specifically pinpoint the exact occurrence of infringement within the vast trove of material being used.
The response also signals that generative AI companies/users cannot bypass existing paywalls for AI training. Although this is a step in the right direction, it raises deeper concerns for the public domain. Copyright owners could begin to lock “public accessibility” behind paywalls to safeguard their interests. This would make it difficult for both human users and AI platforms to access works (even when such access is legally permitted), let alone train on them. This scenario is already common in the publishing industry, where research meant to nourish public knowledge is often inaccessible behind paywalls. Now, this could extend beyond the terrain of academia.
Whether the policy promotes the interests of generative AI companies or those of copyright owners, it is the public domain that stands to suffer. While generative AI platforms are logically expected to grow in the future, the public domain and the interests of creators should not be completely abandoned.
(The writer is an assistant professor at KIIT School of Law)
Disclaimer: The views expressed above are the author's own. They do not necessarily reflect the views of DH.