Behind Govt rollback of advisory on seeking nod for AI services: Industry pushback

Nasscom and the IT Ministry did not respond to a request for comment.

The advisory was simultaneously facing public backlash from some startups in the generative AI space. Aravind Srinivas, founder of Perplexity AI, had called the advisory a “bad move by India”, while Martin Casado, general partner at the US-based investment firm Andreessen Horowitz, had termed the move a “travesty”, which was “anti-innovation” and “anti-public”.

With such adverse feedback, the government changed tack, and in a subsequent advisory on March 15, the IT Ministry made changes recommended by the industry body and AI founders – it scrapped the controversial provision to require government permission before rolling out “under-testing, unreliable” AI services, limited the scope of the guidance to some intermediaries, and removed any mention of companies having to prepare a status report.

In its correspondence, Nasscom said the government’s stated aim of companies offering more information to users about their AI services was “being lost” because of the “approach and drafting” of the advisory.

At the heart of the debate was a tussle between lawmakers and tech companies over the future of safe harbour protections — which afford legal immunity to platforms from the content they host— to generative AI platforms like Gemini and ChatGPT. In India, the tussle began last year, when an AI platform generated responses on a query on various world leaders, including Prime Minister Narendra Modi, which the government felt was uncalled for.

That prompted a broader debate within the government over the lack of India-specific data in the underlying models of these platforms, which can result in responses which may have biases against the country, its politics and history.

However, its solution of asking companies to seek permission before launching systems was met with a pushback from the industry. “The overall objective of the advisory is to empower end users with adequate information so that they make informed decisions while using synthetic media tools. We find this goal is being lost, however, because the approach and drafting of the advisory has sparked concerns and apprehensions across the country,” Nasscom said in its March 7 correspondence.

Demanding that the government “issues a subsequent advisory,” Nasscom said instead of it applying to “all intermediaries,” the guidance “should apply to people making synthetic media tools”.

“…clarify that no explicit permission is mandated from the government of India to make any AI models/software/ algorithm available for public access. Instead, relevant intermediaries will publish reports on capabilities/ limitations of synthetic media tools, provide users, robust controls for providing real-time feedback,” Nasscom submitted.

“Instead of action taken cum status report, the advisory may state that the government of India wishes to convene dialogues with relevant intermediaries with the goal of securing voluntary commitments from them,” the lobby group suggested. The advisory should also clarify that it is “not meant for start-ups publishers, originators, content creators, B2B providers”.

While the updated guidance did not explicitly state this, former Minister of State for IT Rajeev Chandrasekhar had at the time issued the clarification in media statements.