Government finalising rules on labelling AI-generated content, notification soon

The government has completed industry consultations on the proposed mandatory labelling of artificial intelligence (AI)-generated content, and the revised rules are expected to be issued shortly, IT Secretary S Krishnan has said.

Speaking to PTI in an interview, Krishnan said the industry has been “fairly responsible” and understands the rationale behind labelling AI-generated content. He added that there has not been any serious pushback from stakeholders on the proposal.

According to Krishnan, the main feedback from industry players relates to seeking clarity on the extent of modification that would trigger mandatory labelling. Industry participants have sought clear distinctions between substantive or material changes made using AI and routine technical enhancements that do not alter the underlying content.

“Based on the inputs we have received, we are just consulting the other ministries within government, saying that these are the changes which have been suggested… so which changes we accept, which changes we make and what tweaks we make… that portion is on right now, and I think we should come out with the new rules very shortly,” Krishnan said.

On whether the industry is opposed to the move, Krishnan said, “I don’t think they are against it.”

He clarified that the proposed rules do not involve registration, third-party verification, or any form of restriction. “All that is being asked is label the content,” he said, adding that citizens have the right to know whether a piece of content has been generated synthetically or is authentic.

Krishnan pointed out that even minor AI-driven edits can significantly change the meaning of content, while some technical enhancements, such as those carried out by smartphone cameras, may improve quality without altering facts.

“Most of the reaction is the degree and what kind of change because now advanced technology is such that there is some modification or the other in some sense. In some cases, modification can be very small, but that, in itself, can make a difference… One word in a sentence can make a huge difference to what the outcome is,” he said.

He noted that the use of modern devices and technology typically involves some level of enhancement. “The way that you photograph or take a video or record something, the phone itself enhances some of this, tries to make it better,” Krishnan said, adding that the industry wants assurance that such technical improvements, which do not alter substance, are not treated as AI-generated modifications.

“These kind of reasonable asks, we can certainly accommodate,” he said.

However, Krishnan cautioned that excluding all forms of modification could pose challenges. “Even one or two words changing in a particular sequence of conversation could have a completely different effect and impact,” he said, explaining that altering a small part of an otherwise real conversation using AI could change its meaning entirely.

He added that while creativity has its place, transparency remains essential. “People have a right to know this is actual real stuff, and this is not real,” he said.

In October, the government proposed amendments to the IT Rules to mandate clear labelling of AI-generated content and enhance the accountability of large digital platforms such as Facebook and YouTube for verifying and flagging synthetic information. The move aims to curb user harm arising from deepfakes and misinformation.

The IT Ministry has said that deepfake audio, videos and synthetic media going viral on social platforms have shown the potential of generative AI to create “convincing falsehoods,” which can be weaponised to spread misinformation, damage reputations, manipulate or influence elections, or commit financial fraud.

According to the ministry, the proposed amendments provide a clear legal basis for labelling, traceability and accountability related to synthetically generated information. It had invited stakeholder comments on draft rules mandating labelling, visibility and metadata embedding for synthetically generated or modified content to distinguish it from authentic media.

Under the draft rules, platforms would be required to label AI-generated content with prominent markers or identifiers, covering at least 10 per cent of the visual display or the initial 10 per cent of the duration of an audio clip.