Union Finance Minister Nirmala Sitharaman said she has asked the Reserve Bank of India (RBI) whether artificial intelligence (AI) can be regulated with a light touch, after telling a Mumbai audience that a leading AI model is capable of influencing elections without voters realising they are being influenced.

“There are times that I’ve checked up with RBI. Is there a way in which we can do a soft touch regulation? Is there a way in which, without affecting the innovative threshold, is there a way in which we can do something?” she said in the valedictory address at the Global Fintech Fest 2026.

What she said about the model

Sitharaman described a large language model she declined to name as being “fully capable of influencing the opinion in the ground without the ground even sensing that they are being influenced.”

She then framed the concern in her own capacity. “To the extent that me being a minister, being in politics, I’m clearly seeing it capable of influencing elections,” she said.

She was describing a capability rather than alleging an incident, and did not say any election had been affected. Separately in the same passage, she referred to a country being attacked by “several agentic groups together with one group of rogue operators”, which she said had compromised an entire national system. She did not name the country and gave no further detail.

The resignation she pointed to

Sitharaman said public attention had been drawn this week when a researcher resigned from one of the world’s leading AI laboratories, quoting him as saying commercial frontier labs are “racing straight to self-improving superintelligence and gambling with our lives.”

She named neither the researcher nor the laboratory. The quotation matches remarks reported on September 9 by TechCrunch, by a researcher who resigned from Anthropic.

What gave the resignation weight, she said, was not the post itself but the fact that senior figures within the frontier AI ecosystem acknowledged the substance of the warning without naming themselves.

She turned that into a demand. “Who from the global AI industry will stand up and reassure the public? Who will demonstrate that safeguards are actually keeping pace with the speed? Where is the structured collective effort to provide credible and transparent answers?”

Leaving the question of AI going rogue for society at large to answer, she said, is “an unsustainable distribution of risk.”

The precedent she cited

Sitharaman pointed to the European Union’s transparency rules as an example of a jurisdiction acting. She said Article 50 of the EU’s AI Act took effect on August 2 this year, requiring providers to tell people when they are interacting with an AI system, to mark synthetic text, image, audio and video in machine-readable form, and to disclose deepfakes.

She said the penalty is 3% of annual global turnover, and stressed that it applies to worldwide revenue rather than revenue earned in the EU. The Act in fact provides for fines of up to 3% of total worldwide annual turnover, or €15 million, whichever is higher, for breaches of obligations other than the prohibited practices in Article 5.

Where she said industry’s own duty begins

Sitharaman drew a line between what the RBI covers and what it does not. The governor’s assurances on the regulatory sandbox and the innovation hub address the regulated perimeter well, she said. Her remarks were aimed at those operating beyond it — fintechs and broader technology companies.

“Where the regulator’s writ ends, your own standards must begin,” she said.

She also proposed what she called a federated industry platform to bring the Indian technology ecosystem together, not to speak for individual companies but to give the sector a collective voice with foreign regulators “before regulatory positions harden and barriers are erected.” She said it could work on interoperable standards, shared cybersecurity practices and licensing, and help Indian firms commercialise intellectual property across jurisdictions.

In the same breath, she set a reciprocal obligation. Technology companies, she said, “need to be fair to the jurisdictions where they earn their profits”, meeting tax obligations that arise there and giving something back.

On tokenisation, she pushed the RBI on the digital rupee

Sitharaman said REC Limited completed India’s first tokenised corporate bond pilot under the SEBI regulatory sandbox on Monday, with the bond and the digital rupee moving “at the exact same instant.” The pilot was announced on September 8.

What made it possible, she said, was the presence of a central bank digital currency to settle the money leg. “Every tokenisation architecture being tested across the world eventually returns to the foundational question: what is the money leg made of?”

On that basis, she urged the RBI to take its wholesale and retail CBDC pilots further and to “continue to sharpen its capabilities with the digital rupee” — a public prompt from the finance ministry to the central bank on a programme on which the RBI has so far proceeded cautiously.

She also confirmed that the RBI recognised the Unified Fintech Forum as a second self-regulatory organisation for the fintech sector on September 10, as MediaNama reported this week.

Questions MediaNama has sent to the Ministry of Finance and the RBI

• What form of soft-touch AI regulation has been discussed between the ministry and the RBI, and does it extend beyond the financial sector?
• Which authority does the government consider responsible for AI-driven manipulation of electoral opinion?
• Is the government examining transparency and labelling obligations along the lines of Article 50 of the EU’s AI Act?
• Has any proposal for a federated industry platform been put to industry bodies, and who would convene it?
• Following the finance minister’s remarks, does the RBI intend to expand the scope or timeline of its CBDC pilots?

MediaNama has written to the Ministry of Finance and the Reserve Bank of India and will update this story if they respond.

Also read: