Oct 23, 2025

AI content on social media may be labelled to fight rising tide of deepfakes, misinformation ETtech Synopsis India is mandating social media users to declare AI-generated or modified content, with platforms required to implement verification measures. This move aims to combat deepfakes, with disclaimers needing to occupy 10% of visual display or audio duration. The government is also centralizing takedown notices to senior officials. ETtech The government has proposed compulsory declarations by all social media users when posting AI-generated or modified content, with the platforms told to deploy technical measures to verify these. The move is aimed at curbing the rapid rise of AI-based deepfakes on the internet, officials said. Stakeholders have been asked to submit feedback on the latest AI mandate, unveiled on Wednesday, by November 6. There’s no clear timeline for when the rules will take effect, officials said. This initiative builds on earlier IT rule amendments from 2022 and 2023. Electronics and IT minister Ashwini Vaishnaw said the move was necessitated by a deluge of complaints over dangerous deepfakes by citizens, including Parliamentarians. In a key move, at least 10% of the visual display area, or the initial 10% of an audio clip’s duration, will have to be devoted to such disclaimers, the draft amendment to IT rules said. This unique metadata will remain permanently embedded in content to ensure instant identification. Intermediaries are prohibited from allowing the removal or alteration of these labels, it added. ET reported last week that India was planning such a move to stop deepfakes, after studying similar legislation put in place by China and the European Union. ‘Synthetic’ content definition incorporated On Wednesday, officials confirmed the upcoming framework will establish further mechanisms and standards to clearly mark AI-based content. The ministry of electronics and information technology (MeitY) also amended IT rules to ensure only senior government or police officials can ask social media platforms to take down unlawful content. This comes in the backdrop of a court battle between Elon Musk’s X and the Centre. Draft amendments to the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021, published by MeitY on Wednesday include a clear definition of “synthetically generated information” as content artificially or algorithmically created, modified or altered using computer resources in a manner that appears authentic. Officials said this definition is incorporated into existing provisions dealing with unlawful information, making synthetic content liable under the same rules that aim to control other harmful or unlawful materials. ETtech The rule changes will impact all significant social media intermediaries (SSMIs) or those with 5 million or more registered users in India. Google-owned YouTube; Meta-owned Facebook, Instagram, Threads, and WhatsApp; X (formerly Twitter), Snap, LinkedIn, and ShareChat, will now have to obtain user declarations on whether the uploaded content is synthetic, deploy automated tools to verify these declarations, and ensure synthetic content is clearly marked with appropriate labels, or will be considered non-compliant. However, the amendments also clarify legal protections for intermediaries that remove or disable access to synthetic content that’s regarded as harmful or unlawful, reinforcing their safe harbour status under Section 79(2) of the Information Technology Act, 2000. Harbour or deep sea? The rules can dilute safe harbour protection for intermediaries and potentially subject them to criminal liability for their users’ content even in the absence of a government or court notification, said Sachin Dhawan, deputy director, The Dialogue, a tech policy think tank. Akash Karmakar, partner at law firm Panag & Babu, said, “The first step is to label this correctly as a shift of liability for third-party content unto the intermediaries so they are responsible for taking down AI-generated content.” A platform, unlike a publisher of deepfake content, should become complicit in a crime by their passiveness, only where the platform has met a bright-line threshold of complicity or negligence, he said. Ikigai Law partner Aman Taneja said, “The definition of synthetically generated information is very broad and would cover even things like filters, or even a cropped image. The obligation of mandatory visual/audio watermarks on all content generated by AI tools ignores several beneficial use cases where these tools are used.” Google, Meta, X, Snap and ShareChat did not respond to queries. High-level authority In a related move, MeitY has now amended the IT rules to ensure that directions to digital intermediaries including social media platforms, internet service providers and search engines for removing unlawful content, are only issued by certain senior officials. From November 15, only senior government officers at joint secretary levels and above, and law enforcement personnel at deputy inspector general of police rank and above will be able to send out intimations as a result. Secretaries at the home ministry and MeitY will also hold a monthly review of all such notices with the information technology department secretaries of state governments. “The accountability of the government increases with this change, and will be giving a reasoned intimation whenever any such order is passed,” Vaishnaw said. “These will be passed only at a senior level.” Under rule 3 (1)(d) of the IT Act , intermediaries are required to remove unlawful information upon receiving actual knowledge either through a court order or notification from the appropriate government. Safe harbour status is also lifted when an intimation is given to the intermediary about unlawful content and the intermediary fails to remove it. But indiscriminate and inefficient notification by lower-level officials has created complications, officials said. Earlier this month, IT secretary S Krishnan had told central officials that notices under relevant sections of the Information Technology Act 2000 and IT Rules 2021 should be sent in a suitable format and be written in clear language, coupled with legal provisions. Add