Trust N.R. Narayana Murthy? First Check If It's An AI Video

One of the dangers of AI is the advent of convincing deepfakes generated by morphing images, voice notes and videos that can hoodwink you into erroneously revealing financial details or making payment to scamsters

- Shivangini Gupta

AI can read your face, mimic your voice, and rewrite your words in seconds. A few seconds of audio from a WhatsApp note can be used to clone your voice; an old clip from a conference reshaped into a high-definition video of you endorsing a Ponzi scheme, and a stolen profile photo to advertise a fake wealth advisory site. But these aren't crude fakes: the lip-sync is near perfect, the body language convincing.

Among the many dangers of AI is the advent of new forms of fraud and convincing deepfakes that can dent your reputation and rob you of your credentials or investors their money.

Rajesh Krishnamoorthy, an adjunct professor at the National Institute of Securities Markets (NISM) and an independent director in the alternate investments space, who became a victim of a deepfake recently, says that with the advent of AI, now it is very easy to create morphed images and videos.

Cybersecurity professionals say AI-powered scams often employ a mix of stolen material, synthetic voices, and emotional triggers.

Rise In Deepfakes

On a global level, deepfake videos of Ukrainian President Volodymyr Zelenskyy urging soldiers to surrender have surfaced. In India, AI is used in matrimonial frauds, political smear campaigns, blackmail, and to lure investors into fraudulent schemes.

The Reserve Bank of India (RBI) had to issue a public advisory in November 2024 when deepfake videos began circulating showing the RBI governor pitching investment schemes. RBI said these videos were fake, its officials do not endorse or advise on such schemes, and the public should ignore them.

The government's cybersecurity agency, CERT-In, also sounded alarms about deepfake-based fraud and the risk of generative AI. It issued an advisory in November 2024 highlighting how Al-generated media was being weaponised for impersonation scams, disinformation, and social engineering attacks.