Half of Indians watched multiple fake videos in last year; 7 in 10 want deepfake creation to be a criminal offence: Survey

As the Centre prepares a dedicated law to criminalise deepfakes, a new LocalCircles survey suggests public opinion is firmly behind tougher action. Nearly seven in ten Indians surveyed believe creating fake or deepfake videos should be treated as a criminal offence, while almost half said more than one in ten videos they watched over the past year later turned out to be fake.

The survey, which received over 35,000 responses from citizens across 336 districts, found that 68% of respondents favoured making the creation of fake or deepfake videos a criminal offence. Another 14% said such content should attract a heavy financial penalty, while 9% preferred a milder punishment and another 9% were undecided.

“The urgency is backed by hard numbers,” the report said, pointing to the rapid rise of deepfakes and the increasing financial and social harm caused by AI-generated content.

The findings come as the government informed the Supreme Court that it is preparing a dedicated Bill to define AI-generated deepfakes and digital arrest scams as standalone criminal offences. The proposed legislation is expected to build on the IT Rules amended in February this year, which require intermediaries to label AI-generated media and remove unlawful synthetic content within a specified timeframe.

According to the survey, 48% of respondents said over 10% of the videos they watched on social media, OTT platforms or WhatsApp in the last 12 months were later found to be fake. Breaking that down, 24% said more than half the videos they watched later turned out to be fake, 12% said between 25% and 50%, another 12% said between 10% and 25%, while 28% estimated the figure at up to 10%. Sixteen per cent said they had rarely encountered fake videos and 8% did not provide a clear response.

“Created using AI that can convincingly superimpose faces, clone voices and fabricate realistic audio-visual content, deepfakes now go well beyond entertainment,” the report noted. “They are used to spread misinformation, damage reputations, enable financial fraud and manipulate public discourse.”

The report also cited broader industry estimates to underline the scale of the challenge. It said the number of deepfake files globally has risen from around 500,000 in 2023 to nearly 8 million in 2025. It also referred to the Thales 2026 Data Threat Report, which found that 65% of organisations surveyed had experienced deepfake-driven attacks, while 55% reported damage from AI-generated misinformation. Separately, a McAfee study estimated that about 90% of Indians had encountered fake or AI-generated celebrity endorsements, with affected victims losing an average of ₹34,500.

Heightened scrutiny calls for stricter laws

The survey comes at a time of heightened scrutiny of technology platforms. On August 5, the Parliamentary Standing Committee on Communications and Information Technology questioned executives from Meta, Google, X and Snap over compliance with India’s deepfake labelling rules. The report said the CEO of the Indian Cyber Crime Coordination Centre (I4C) informed the committee that none of the four platforms was complying with the statutory obligation to label synthetically generated information.

“With a dedicated Bill on standalone criminal offences under preparation, clear statutory recognition of deepfake creation as a crime would strengthen investigations, improve conviction rates, and send an unambiguous signal that such acts constitute serious criminal conduct rather than harmless spoof,” the report said.