The draft guidelines propose that AI-generated visuals carry labels covering at least 10 per cent of the screen or duration to curb deepfakes.
Social media firms must ensure metadata traceability, user disclosure, and automated checks for AI-made material.
The IT ministry has sought public feedback by November 6 as courts hear celebrity deepfake cases, including one by Abhishek and Aishwarya Rai Bachchan.
India Moves to Mandate Labelling of AI-Generated Content to Curb Deepfakes
New draft rules require tech giants like OpenAI, Meta, and Google to visibly tag AI-made visuals and audio, citing rising misinformation risks.
Summary
Prompted by similar actions taken by China and the European Union, India's government proposed on Wednesday that social media and artificial intelligence companies prominently mark content produced by AI to combat the spread of misinformation and deepfakes.
With almost 1 billion internet users, the stakes are high in a nation with a diverse population of ethnic and religious groups. AI deepfake films have scared authorities before elections, and fake news has the potential to incite violent conflict.
The new guidelines place additional duties on companies like OpenAI, Meta, X, and Google by requiring platforms to label AI-generated material with markers that span at least 10 per cent of the surface area of a visual display or the first 10 per cent of the duration of an audio recording.
According to the Indian government's draft plan, social media businesses would also need to adopt appropriate technical mechanisms to provide checks and balances and get a user declaration on whether the submitted information is AI-generated.
India's IT ministry stated that the regulations will "ensure visible labelling, metadata traceability, and transparency for all public-facing AI-generated media," and it is asking the public and industry to provide feedback by November 6.
"To cause user harm, spread misinformation, manipulate elections, or impersonate individuals, the potential for generative AI tools to be misused has grown significantly," the ministry said.
The rules about covering 10 per cent of surface area are "among the first explicit attempts globally to prescribe a quantifiable visibility standard," said Dhruv Garg, founding partner at public policy research firm, Indian Governance and Policy Project.
High-profile deepfake litigation is being heard in Indian courts. Bollywood actors Abhishek Bachchan and his spouse, Aishwarya Rai Bachchan, have challenged YouTube's AI training policy and urged a judge in New Delhi to ban the production of AI videos that violate their intellectual property rights.
According to him, if the regulations are put into effect, AI platforms in India will have to have automated labelling systems to recognise and mark AI-generated content at the time of creation.
Advertisement
India is emerging as a big market for AI firms.
With inputs Reuters.
Published At: