3 min readJul 29, 2026 06:20 AM IST
First published on: Jul 29, 2026 at 06:20 AM IST
There are three takeaways in the the Delhi High Court’s recent decision in ANI v. OpenAI. The first is interpreting copyright law in a technologically neutral manner. OpenAI contended that its activities qualified as research, an activity deemed non-infringing under the law. ANI’s position, however, effectively confined the protection accorded to research to activities undertaken by human beings. The court rejected this distinction, observing that machines are increasingly capable of undertaking tasks that were once the sole remit of humans, including research. To deny protection simply because such activities are machine-mediated would risk of stymying societal progress when these technologies ultimately operate at the behest of and for the benefit of humans.
By drawing parity between machine- and human-led activities, the court implicitly invokes technological neutrality — a core tenet of technology governance, which holds that the legal position of an activity should not change merely because technology alters the means through which it is undertaken. In the context of AI and copyright, this matters because research, analysis, and learning should not lose protections because humans are being assisted by machines. Otherwise, every technological advance could turn laws intended to encourage creativity and learning into barriers for innovation.
The second is the court’s acknowledgement of dynamic ties between the technology and copyright industries. There is considerable evidence that AI is already augmenting creative activity. According to a study by Adobe, 83 per cent of creators now use AI for their work. A 2025 report by the London School of Economics finds that AI use in the creative industry has sparked dynamism – different sub-sectors have developed tailored tools, often built by start-ups. This is the kind of Cambrian explosion of innovation that lies at the heart of copyright law’s purpose.
The third is: When trying to reconcile the interests of AI developers and creators, begin by asking whether the particular use, in this case, AI training, is infringing at all, rather than presuming there is a compensable claim. Earlier this year, a committee constituted by the Department for Promotion of Industry and Internal Trade proposed a direction on copyright and AI training that takes a position in favour of the latter. It proposed a licensing regime under which AI developers could use copyrighted works for training subject to compensation for rightsholders. This regime could raise the cost of developing models, with the burden falling particularly heavily on start-ups and smaller developers. And as AI tools become embedded in creative workflows, increasing the cost and uncertainty surrounding their development can also affect creators and creative businesses.
Rather than choosing between copyright owners and AI developers, the DPIIT panel must work to preserve the conditions under which creative production and technological innovation can flourish. Treating their interests as antagonistic risks producing the outcome copyright law is intended to avoid: Protecting one form of creativity by constraining the emergence of another.
The writer is director of the Esya Centre, a tech policy think tank. Views are personal
