PM Modi flags deepfake threat, cites his garba video

Aditi Tandon

New Delhi, November 17

Prime Minister Narendra Modi on Friday said deepfake videos posed a grave danger to society, revealing for the first time that he had asked ChatGPT makers to red-flag the hazards of artificial intelligence (AI)-generated videos while putting these out.

WHAT ARE DEEPFAKES

A type of AI used to create convincing image, audio and video hoax. These involve manipulation of facial appearance through deep generative methods.

Addressing a Diwali get-together at the BJP headquarters here, Prime Minister Modi said he recently saw a video of him performing garba even though he had not done so since schooldays. “There is this video where I am shown doing garba. It is even being circulated,” the PM said, signalling India’s strong position on the harmful potential of AI and the need for stringent regulation.

Govt advice to intermediaries like FB, Twitter

• Ensure diligence and reasonable efforts to identify misinformation and deepfakes
• When detected, act expeditiously to remove deepfakes under IT Rules, 2021, and ensure users are not allowed to host such content
• Remove fake content within 36 hours of reporting

Urging media’s support to educate the masses on the ills of technology, the Prime Minister said the AI-generated deepfakes could pose a huge challenge in a diverse society like India’s.

“Just like tobacco packages mention the product is hazardous/cancerous, I told ChatGPT (free to use AI system) team during my last meeting with them that they must indicate deepfake videos as such so that people know these are not real,” the PM said amid growing outrage over deepfakes involving actors Rashmika Mandanna, Katrina Kaif and Kajol.

Following Rashmika Mandanna’s deepfake which went viral, the government had issued an advisory to all significant social media intermediaries, asking them to exercise due diligence and make reasonable efforts to detect misinformation and deepfakes and take them down.

The Ministry of Electronics and Information Technology, in its advice, said intermediaries must remove deepfakes within 36 hours of reporting or face legal action.

Current regulations around the AI, however, remain inadequate with no standalone AI regulation law in place. Under the existing Information Technology Act, Section 66 D deals with punishment for cheating by impersonating someone using a computer resource.

It says anyone found guilty of impersonation using communication devices or resources may face imprisonment for up to three years and a fine of up to Rs 1 lakh.