Artificial Intelligence is rewriting the economics of cybercrime, and the change is arriving faster than institutions can respond. Financial crime specialists now warn that a single fraudster, armed with off-the-shelf AI tools, could soon run as many as 10,000 automated scams in a single day. What once demanded call centres, scripts and manpower has been compressed into something one person can operate from a laptop.
An “Industrial-Scale” Criminal Enterprise
Dr Shlomit Wagman, former head of Israel’s Financial Intelligence Unit and founder of Integria AI, has cautioned that the pace of AI adoption among criminal networks could trigger a global “tsunami” of financial fraud. Fraud, she argues, has shifted from a manual activity into a machine-driven one, with AI systems capable of generating thousands of tailored, convincing messages, voices and videos within seconds.
That shift has already produced real losses. Fraudsters used AI voice-cloning technology to steal roughly ₹2.2 crore from a UK-based energy company by mimicking an executive’s voice on a phone call. In a more elaborate case, criminals staged a deepfake video conference to defraud a Hong Kong-based multinational of close to ₹215 crore, with employees authorising transfers after “seeing” and “hearing” colleagues who did not exist.
India’s Own Reckoning With AI Fraud
For Indian readers, this is not a distant risk. The Ministry of Home Affairs has told Parliament that Indians lost more than ₹22,495 crore to cyber fraud in 2025, with the Indian Cyber Crime Coordination Centre projecting the true figure, including unreported cases, could approach ₹1.2 lakh crore. Over 2.81 million complaints were filed last year, translating to more than 7,700 cases a day.
Deepfakes and voice cloning form a growing share of that toll. A 2025 survey by the Observer Research Foundation found that 47 per cent of Indian adults had personally been defrauded by, or knew someone defrauded by, an AI voice-cloning or deepfake scam — nearly double the global average. Fraudsters have deepfaked business leaders and celebrities to endorse fake trading platforms, and “digital arrest” scams using AI-generated faces of police or CBI officials have drawn more than 90,000 complaints, with the Supreme Court estimating losses near ₹3,000 crore.
Prof Triveni Singh, a cybercrime expert and former IPS officer, has warned that AI now allows fraudsters to send highly personalised calls, messages and videos to thousands of potential victims within minutes. He stresses that technology alone cannot solve the problem — digital awareness, identity verification and independent confirmation of any financial request remain equally critical.
Regulators and Banks Face a Widening Gap
The Future Crime Research Foundation has cautioned that AI-generated voices, videos and identities are now realistic enough to defeat traditional verification, and has urged financial institutions to adopt multi-factor authentication, AI-powered fraud detection and regular staff training. India’s own institutional response — the I4C, CERT-In, the 1930 helpline and new IT Rules requiring platforms to act on flagged synthetic content within three hours of an order — reflects the urgency, even as recovery rates for stolen funds remain around a quarter of losses.
Banks, regulators and law enforcement agree that policing alone will not be enough. Real-time transaction monitoring, behavioural analytics and stronger identity checks are being positioned as the next line of defence, alongside a public that is learning, often the hard way, to distrust a familiar voice on the phone.
Experts advise a simple discipline: never transfer money on the strength of a call, video call or message alone. Any request involving urgency, secrecy or an AI-generated voice or video should be independently verified through a separate, trusted channel before any payment is made. As AI adoption deepens across both legitimate finance and organised crime, that habit of verification may prove to be the most durable safeguard available.
