The government move to introduce an artificial intelligence law that will safeguard the commercial and creative interests of content creators including news publishers will help balance the interests of publishers with the needs of those developing generative AI (GenAI) models, legal experts said.
Union IT minister Ashwini Vaishnaw had told ET last week that the government is looking at introducing such an AI law.
Legal experts said it is a “timely” move amid rapid expansion of AI-generated content that could significantly impact publishing industry and has led to concerns over copyright infringements.
“Specific legislation on this issue will help India find the appropriate balance between enabling technology innovation (through AI) while valuing and protecting news reporting, and artistic innovation,” said Arun Prabhu, partner and head – technology & telecommunication at law firm Cyril Amarchand Mangaldas.
Regulating the use of news and content by AI companies is “a global regulatory and policy hotspot” and India’s law can address whether AI platforms can circumvent paywalls, what is considered ‘fair’ use, how direct and indirect benefits should be shared, credit and attribution, etc., Prabhu told ET.
Vaishnaw had told ET that the government through the new law seeks to secure the rights and sharing of the proceeds among news publishers, content creators and AI-enabled technologies, “while keeping good space for innovation”.
The proposed law could be in the form of separate legislation or part of the Digital India Bill that will replace the 24-year-old IT Act, he had said.
Ranjana Adhikari, partner at IndusLaw, said, “The interface of generative AI models with copyright law is a recent phenomenon and Indian jurisprudence has not evolved sufficiently to address the issue.”
She noted that globally, multiple suits have been instituted before courts by prominent publishers specifically on this issue. A case in point is the New York Times vs OpenAI case where adjudication is awaited.
The recently passed EU AI Act requires AI models to comply with EU copyright law and also publish detailed summaries of the content used for training.
“In my view, any future Indian law (whether as a part of the Digital India Bill or a standalone statute), which attempts to safeguard the IP and commercial interest of publishers in the age of AI, will have to be preceded by an extensive exercise which firstly identifies the various ways in which AI models use existing copyrighted material for their training and secondly analyses how much of such use legitimately falls within the domain of infringement,” Adhikari said.
Under the existing provisions of the Indian Copyright Act, 1957, to avoid infringement while using any copyrighted work, one has to either fall within one of the exceptions under its Section 52 (such as fair dealing) or procure an appropriate license or claim other legal defences (as applicable) such as de minimis usage.
“Once this background exercise is done, the ideal approach could then be to enact a law which codifies certain definitive principles and safeguards, specifically in the context of AI training and copyright law,” Adhikari said. “This may then lead to publishers and AI models arriving at mutually beneficial arrangements for use of copyrighted material.”
Any such law should, ideally, not be too prescriptive or regressive and should not lose sight of the transformative potential of generative AI, she said.
Shahana Chatterji, partner at Shardul Amarchand Mangaldas & Co, said consultation for a new AI framework is a welcome move.
“It is an interesting point at which we are poised,” she said. “There are existing regulatory frameworks that can continue to apply to risks created from AI solutions, such as some of the offences established under the IPC (and the upcoming Bharatiya Nyaya Sanhita framework). On the other hand, there is an absence of a uniform legislation that regulates the manner in which AI solutions are developed and deployed.”
The conventional liability frameworks that exist for the digital sector, such as intermediary liability, does not fit the AI ecosystem which is significantly more complex than an intermediary based ecosystem, Chatterji said. “The logical next step would be to initiate stakeholder consultation and therefore what the government is planning is a welcome move,” she said.
Chatterji said while it is too early to comment on the exact contours of such a framework, she looks forward to “discussions on the complexity of developing such solutions, the multiple stakeholders involved in developing and deploying these solutions, the exponential impact it can have on the growth of the Indian economy, and the manner in which potential harms can be managed and mitigated”.
Imposing obligations that increase the cost of doing business or discourage the use of Indian content by AI makers could deal a blow to India’s ‘innovation economy’, experts cautioned.
