Govt’s AI advisory sparks expert discontent, urges clarity

The government has recently clarified its advisory on Artificial Intelligence (AI), specifying that it only applies to large firms and not startups. However, experts remain unsatisfied, citing the government’s March 1 advisory as unclear and overly broad.

A recent paper by the Global Trade Research Initiative (GTRI) points out the ambiguity in the advisory and suggests that the government make some changes. It suggests that the advisory should clearly distinguish between ‘significant platforms’ and startups or smaller companies.

In the paper released on Thursday, the research group argued that since AI is a new and emerging field, every firm could be a startup, and even large platforms could hive off their AI venture as a fresh startup.

“The AI advisory and subsequent clarifications have not sufficiently clarified to whom the advisory is meant to apply. For example, what constitutes a ‘startup’ that is exempted from the requirements under this advisory,” said Sidharth Deb, Public Policy Manager at TQH Consulting.

Further, the mandate to take approvals before deploying AI models or platforms could stifle innovation in India’s emerging technologies sector, said experts.

According to a recent report by Nasscom-BCG, India’s AI market, growing at a compounded annual growth rate (CAGR) of 25 to 35 per cent, is projected to reach $17 billion by 2027. As the investments in AI continue to increase, the demand for AI talent in India is also expected to grow at a CAGR of 15 per cent till 2027, the report said.

Regulatory ambiguity
“The advisory introduces a ‘permission prior to deployment’ process for under-tested and unreliable AI platforms. The absence of details around the process of seeking such permission by the government and the timelines involved remain unclear,” said Jameela Sahiba, Senior Programme Manager, Emerging Technologies, The Dialogue.

“This fragmented regulatory effort may result in a less conducive environment for innovation, as players may struggle to comply with unclear directives, ultimately impacting the growth and sustainability of the AI ecosystem in India,” she said.

As a step to regulate AI, experts say that the government should pursue a comprehensive public consultation on the issue, and avoid issuing advisories. “Public consultation with AI regulation will help marshall collective intelligence where industry and experts can identify risk-based safeguards and standards which are consistent with the idea of responsible innovation,” said Deb.

Chatbot fiasco
Recently, Google’s chatbot – Gemini was criticised for allegedly producing biased responses to a user’s questions about PM Modi. Union Minister Rajeev Chandrasekhar had said that the responses by Gemini were a direct violation of IT rules, as well as several provisions of the Indian criminal code.

While clarifying the advisory, Chandrasekhar in a post on X said, “Process of seeking permission, labelling and consent-based disclosure to users about untested platforms is an insurance policy to platforms who can otherwise be sued by consumers.”

MeitY has issued the advisory to intermediaries and AI platforms and asked them to submit an status-cum-action-taken report within 15 days, around mid-march. Business Standard