To tackle AI-generated deepfakes and harmful synthetic content, the Central government has approved 13 ‘Responsible AI’ projects focused on developing deepfake detection technologies. The Ministry of Electronics and Information Technology of India (Meity) has also announced stricter rules for social media platforms such as Meta and X to protect users and stop the spread of fake AI-generated content.
The announcement was made by Union Minister of State for Electronics and Information Technology Jitin Prasada in the Lok Sabha. According to him, the growing use of AI has made it easier to create fake videos, images and audio that can be used for scams, impersonation, misinformation and other cybercrimes.
To address these risks, the government has strengthened existing laws and amended the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021.
One of the biggest changes is that online platforms will now have to act much faster against unlawful AI-generated content. If they receive a valid government or court order, they must remove such content within three hours, instead of the earlier 36-hour deadline.
The government has also shortened the timelines for handling user complaints, especially in sensitive cases like nudity, impersonation and other harmful content.
The amended rules also require intermediaries to clearly label AI-generated content and maintain traceable metadata wherever such content is allowed. This is expected to help users identify synthetic content more easily and reduce the chances of people being misled by deepfakes.
Social media companies have also been asked to deploy suitable technical measures, including automated tools, to detect and prevent the spread of unlawful AI-generated content.
Alongside the regulatory changes, the government has approved 13 ‘Responsible AI ‘projects across educational institutions to improve deepfake detection technologies.
Among the key initiatives is Saakshya, a multi-agent deepfake detection framework being developed by IIT Jodhpur and IIT Madras. The system is designed to identify AI-generated fake videos and images with improved accuracy.
Another project, AI Vishleshak, is designed to detect fake audio and video. It uses AI to analyse media files and identify whether they have been manipulated or altered.
Meanwhile, IIT Kharagpur is working on a Real-Time Voice Deepfake Detection System that can identify AI-generated voices during live conversations, making it useful in tackling voice scams, financial fraud and identity impersonation.
According to the government, these research projects will help build homegrown AI solutions capable of detecting increasingly sophisticated deepfakes.