Overview
- MeitY’s draft would require platforms to disclose synthetically generated information with visible on‑screen or initial‑audio labels, plus watermarking, metadata identifiers, user declarations and automated verification by Significant Social Media Intermediaries.
- Government sources signaled a possible reduction of the visible‑label coverage from 10% to roughly 5–7%, while MeitY’s secretary said the ministry will consider stakeholder submissions before a final decision.
- Industry groups warned that large labels could degrade user experience and raise costs, and urged a risk‑based approach with carve‑outs for non‑deceptive or B2B uses and reliance on durable machine‑readable provenance.
- Civil‑society advocates cautioned that broad SGI definitions could compel speech and drive general monitoring, noting that labels and metadata can be stripped or evaded and automated detection has non‑trivial error rates.
- Major platforms have begun limited tools for disclosure and detection—such as YouTube’s likeness detection and Meta’s labelling—while legal analysts argue transparency alone will not deter AI‑enabled crimes already prosecutable under existing statutes.