SC rules push for deepfake labelling — but is it enough to curb AI misinformation?

The Supreme Court’s growing scrutiny of AI-generated content and the government’s decision to tighten the Information Technology Rules, 2021, have put deepfakes firmly in the spotlight.

In the ruling, social media platforms such as YouTube, Instagram, and X will have just three hours to take down flagged synthetic content once notified by authorities.

They must also label all AI-generated media and embed permanent identifiers or metadata marking it as artificial. The new rules will be implemented from February 20 onwards.

It’s a significant step towards reining in the misuse of generative AI, from fabricated political videos to fake legal citation, but questions remain: is this enough to restore public trust?

More from Tech

Is labelling enough?

For many experts, the new labelling rule is an overdue but incomplete response. It signals recognition that synthetic media has become a mainstream threat, capable of undermining credibility, fuelling misinformation, and damaging reputations within hours.

“India’s move to explicitly bring synthetically generated information within the IT Rules is a significant regulatory step,” said Malcolm Gomes, Chief Operating Officer at Privy by IDfy. “It recognises that deepfakes and AI-generated media are no longer fringe concerns but real risks to public trust and institutional credibility. In my view, this acknowledgement is overdue, because the scale and sophistication of synthetic media today far outpaces the current guardrails.”

However, Gomes warns that labelling cannot be the only safeguard. “Synthetic content spreads extremely fast across platforms, and by the time it is labelled, it may already have influenced opinion, disrupted markets, or caused reputational harm,” he said.

Gomes emphesised that “The real safeguard has to sit much earlier in the process, through proactive deepfake detection tools, stronger content verification systems, and continuous monitoring that can identify manipulated media before it spreads widely. Without this technical backbone, labelling risks becoming a compliance formality rather than a meaningful control.”

Legal experts share this view, calling for a deeper framework that combines judicial oversight with legislative muscle. Atul Kumar, Supreme Court lawyer, cautioned that while the recent steps are welcome, piecemeal measures will not suffice.

“AI is a double-edged sword; if not used cautiously and genuinely, it can inflict injury on the unintended also,” Kumar said.

He pointed out how the courts themselves have begun to encounter the perils of AI misuse. “If AI starts dominating the proceedings of the courts, it would have dangerously intimidating consequences. In Deepak Raheja vs Omkara, AI generated non-existent case references in a rejoinder affidavit,” he said.

“Similarly, the Supreme Court questioned the Election Commission’s use of AI-driven software, observing that such tools were not based on ground realities.”

Kumar said that while the judiciary can take assistance from AI in administrative matters such as case listing, decision-making must remain human-led. “The law is a dynamic concept; only a human is competent to interpret it within context and circumstance,” he said.

“One directive from a court or a government order will not solve the deepfake challenge, we need detailed guidelines and stricter laws.”

Why AI governance is becoming a board-level trust imperative?

Beyond regulation, experts say organisations themselves must confront how AI-generated media could expose them to reputational and legal risks. Deepfakes, once a distant threat, now pose enterprise-level dangers, from manipulated brand videos to fraudulent employee communications.

“AI governance is quickly becoming a board-level trust priority, whether organisations are ready for it or not,” said Gomes. “The misuse of AI-generated media can create regulatory exposure, financial liability, operational disruption, and long-term brand damage. This is not just a content moderation issue, it is an enterprise risk issue.”

He added that corporate boards now need clear visibility into how AI systems are deployed, how authenticity is verified, and how third-party AI risks are managed. “In the years ahead, credibility will depend not just on innovation, but on responsible oversight,” Gomes said.

“Organisations that treat AI governance as a core element of continuous risk management, rather than a one-time compliance exercise, will be far better positioned to retain trust in an environment where manipulation is becoming cheaper and more scalable by the day.”

As deepfakes grow more convincing and accessible, India’s new labelling and takedown rules are a crucial foundation, but far from a solution.

Technology experts and lawyers alike agree that AI governance must evolve faster than AI deception. Without proactive detection, rigorous oversight, and stronger laws, labelling may end up being little more than a digital warning tag on an ever-expanding web of synthetic lies.