Government of India has formally subjected AI generated material like deepfake videos, synthetic audio and manipulated visuals to a formal regulatory structure to the first time. The relocation is by an amendment to information technology intermediary guidelines and digital media ethics code rules 2026, which was notified by gazette notification G.S.R. 120(E).
The new regulations, with the signature of the Joint Secretary Ajit Kumar will be effective since February 20. Under the modified regulations, social networks and online intermediaries will have to identify all regularly generated data by marking them with clear labels that will allow users to recognise them as AI-generated content at first sight.
The government has also enforced such content to be embedded with persistent metadata and unique identifiers by the platforms. These identifiers have to enable tracing of the content to the source and can never be deleted, changed or concealed once applied.
This is the first government definition of information synthetically generated. It involves any audio or visual or audiovisual material produced or substantially modified with the help of computer facilities in a manner that it presents as real and may mislead observers that those people or occurrences they see are real.
But the editing of images as routine as colour correction, compression, translation or noise reduction, is not going to be classified as such as long as the original meaning is not distorted. Training materials, research papers, presentations, illustrative and draft materials based on hypothetical content have also been excluded from the scope of the new definition.
Key social media systems such as Instagram, YouTube and Facebook will experience a higher degree of compliance. Users should be asked to indicate whether their posts are AI generated before posting them.
The platforms must also provide automated tools to prove such declarations through the format, origin and type of the content. In case it is found to be a synthetic one, then the contents should have a visible disclosure label. The lack of action on the violating content can be considered as a lapse of the due diligence.
The government has minimized the time used by the intermediaries in responding to legal orders. Some takedown requests have to be resolved within three hours rather than 36 hours. The previous 15 days have been narrowed down to seven days and 24 hour time has been narrowed to 12 hours.
The revised regulations also associate synthetic content with criminal provisions as per the Bharatiya Nyaya Sanhita, the Protection of Children from Sexual Offences Act as well as the Explosive Substances Act in situations of child abuse material, obscene material, false electronic record, explosive material related or a deepfake that requires a resemblance of the identity or voice of a real person.
Also Read: Reliance Industries Acquires Three Luxury Flats at One Altamount Road for ₹85 Crore
The platforms must also notify the users at least once every three months regarding the punishment concerning the abuse of AI generated content.
Meanwhile, the government has made it clear that intermediaries who operate in accordance with these rules will still be afforded protection under Section 79 of the Information Technology Act under the provision of safe harbour.
