With the sharp rise in AI-generated content on social media, the government has introduced new rules to curb its misuse. This includes deepfakes, non-consensual intimate images, obscene or sexually explicit material, and fake documents created using artificial intelligence tools. The Ministry of Electronics and Information Technology (MeitY) on Tuesday notified amendments to the Information Technology Rules, 2021, which will come into force from February 20, 2026.
Under the new framework, uploading or sharing harmful AI-generated content will be treated as illegal, just like other forms of unlawful online material. Users who violate the rules may face criminal action under existing laws, including the Bharatiya Nyaya Sanhita, 2023. The amendments also place tighter obligations on social media platforms, shorten timelines for removing harmful content, and clarify that platforms will not lose legal protection for taking down such material.
Why govt tightened the rules
According to the government, the amendments were necessary due to rapid advances in artificial intelligence and machine learning. These technologies have made it much easier to create highly realistic synthetic content that can closely mimic real people, voices and events. AI tools can now generate or manipulate audio, images and videos in ways that appear authentic, making it difficult for users to distinguish between real and fake content.
Such content can be used to mislead people, damage reputations, spread misinformation, commit fraud or violate privacy. The government said the new rules aim to ensure that AI-generated content of this nature is regulated in the same way as other illegal online content, without allowing technology to be used as a shield against accountability.
Synthetically generated info
The 2026 amendments focus mainly on a category called Synthetically Generated Information (SGI). As defined under the amended IT Rules, SGI refers only to artificial or algorithmically created or altered audio, visual or audio-visual content that appears real and may be mistaken for genuine people or events.
These include deepfake videos, AI-generated or AI-altered images, and cloned or synthetic voice recordings. Importantly, text-only content is not classified as SGI under this definition. However, the rules clarify that if synthetic audio or visual content is used to commit an unlawful act, it will still be treated as “information” under relevant legal provisions.
This ensures legal implications for deepfakes or cloned audio shared along with captions, descriptions or posts. For instance, a deepfake video shared with a misleading caption, or a cloned voice recording falsely impersonating an official, will be treated as SGI. A chatbot-generated text article without any synthetic audio, image or video will not fall under SGI, though it may still be illegal depending on its content.
Mandatory labelling of AI-generated content
Under the new regulations, lawful AI-generated content must carry clear and prominent labelling to ensure users are not misled. Visual and audiovisual content, such as images or videos, must display a visible label within the content itself stating that it is “synthetically generated.” This label cannot be hidden in captions, descriptions or metadata and must be easily noticeable while viewing the content.
For audio-only content, a clear disclosure such as “This audio is synthetically generated” must be played at the beginning of the recording. Wherever technically feasible, such content must also include permanent embedded metadata or technical provenance mechanisms to support traceability and prevent tampering.
What the rules do not affect
The amendments make it clear that routine, good-faith uses of technology will not be treated as SGI. This includes basic editing, educational or illustrative material, and accessibility improvements, as long as these do not materially alter the original content or create false documents or records.
Lawful uses such as clearly labelled satire or creative synthetic works may also be allowed, provided they do not violate any law and are properly declared as synthetic. However, AI-generated fake certificates, forged identity documents, fabricated official letters or false electronic records are explicitly excluded from these protections and may be treated as unlawful SGI or false records.
Stronger duties for platforms and faster takedowns
The amendments significantly strengthen due diligence obligations under Rule 3 of the IT Rules, 2021. Platforms must now inform users about their obligations every three months instead of once a year. Users must be warned that sharing illegal content, including deepfakes and non-consensual intimate images, can attract legal consequences and, where required, reporting to authorities.
Platforms that enable the creation or sharing of synthetic content must also issue special warnings during onboarding and periodically thereafter, cautioning users against misuse and explaining possible penalties.
Timelines for removing unlawful content have been sharply reduced. Platforms must now remove or disable access to illegal content within three hours of receiving a court order or a reasoned notice from an authorised government officer, compared with 36 hours earlier. Grievance redressal timelines have also been shortened, with especially fast action required in cases involving sexual content, morphed images or impersonation.