India Mulls New AI Law to Safeguard News Publishers & Content Creators

The Indian government is considering the introduction of a new law for artificial intelligence that aims to protect news publishers and content creators while minimising harm to users, as revealed by Ashwini Vaishnaw, India's union minister for electronics and information technology, in an interview with the Economic Times.

Vaishnaw emphasised that the new law will strike a balance and be "strong on securing the rights and sharing the proceeds" among news publishers, content creators, and AI-enabled technologies like large language models (LLMs). He added that the law will provide ample room for innovation.

The proposed legislation could either be an independent law or incorporated into the Digital India Bill, which is set to replace the Information Technology Act, 2000.

"There is a transition happening. Our position is that the transition should not be disruptive because lakhs of livelihoods are involved," Vaishnaw said in the interview.

"Secondly, creativity has to be respected both in terms of intellectual property as well as financial and commercial implications. We have shared these views with tech players. More or less they are in agreement in principle," he added.

"One thought is to form a self-regulatory body. But we don't think that would be enough. We think that this regulation should be done by legislative method. We have already consulted the industry. After elections, we will launch a formal consultation process and move towards legislation," Vaishnaw concluded.

This development comes amid growing concerns over the use of copyrighted material to train generative AI models. In the US, the New York Times sued Microsoft and OpenAI, alleging that millions of its copyrighted articles were used to train OpenAI's generative AI models and ChatGPT. Google was also fined $270 million in France for using news articles to train its generative AI model called Gemini.

Recently, OpenAI CTO Mira Murati could not conclusively confirm whether its new Sora video generation system was trained using YouTube videos. Neal Mohan, the CEO of YouTube, stated that if Sora was trained without the consent of creators, it would violate YouTube's terms of service. He added that Gemini does use some YouTube content, but with the support of its terms of service and individual contracts with creators.

Similarly, Apple is investing millions of dollars to license content from US news organisations to train its generative AI system for the iPhone. Following the New York Times lawsuit, several authors, computer programmers, and musicians have taken legal action against tech giants developing generative AI technology based on large language models, seeking compensation for their work.

News publishers in India are also seeking similar relief through changes to the information technology rules to ensure fair compensation for the use of their content by generative AI models. The Digital News Publishers Association has sent a letter and made representations to the ministries of electronics and information technology, as well as information and broadcasting, to seek protection from copyright violations by generative AI developers.

In Europe, the recently passed AI Act introduces strict guardrails for developers of high-risk AI systems. It mandates that AI models must comply with EU copyright law and provide detailed summaries of the data used to train the models.