The Financial Conduct Authority (FCA) has today announced emergency measures to combat a sharp rise in artificial intelligence-powered fraud schemes targeting British savers. The regulator revealed that AI-generated voice clones and deepfake videos are now being used to impersonate bank staff and financial advisors, tricking victims into transferring life savings into fraudulent accounts. Since the start of the year, losses from such scams have exceeded £300 million, with an estimated 150,000 Britons affected.
In a statement, FCA Chief Executive Nikhil Rathi said the escalation represents a ‘new frontier’ in financial crime, one where AI tools allow fraudsters to operate at scale with unprecedented realism. ‘We are seeing highly convincing synthetic identities and voice mimics that even trained staff struggle to spot,’ Rathi noted. In response, the regulator is activating what it calls ‘Enhanced Protection Protocols’, a three-pronged strategy involving forced cooling-off periods for suspicious transactions, mandatory real-time verification checks for new payees, and a central database of verified account holders.
For ordinary savers, these changes mean that high-value transfers over £1,000 may now be delayed by up to 48 hours while banks cross-check audio and biometric data against known patterns. The FCA is also working with telecoms and tech firms to identify and block AI-generated phone numbers that mimic official bank lines. However, critics warn that these fixes may be too little too late. Privacy campaigners have raised concerns about the centralised database, calling it a ‘backdoor surveillance tool’. Yet Rathi dismissed those fears, insisting that ‘security must take precedence in this emergency’.
Behind the scenes, the fraud surge exploits a vulnerability inherent in our digital infrastructure: the ease with which generative AI can produce believable synthetic media. Fraudsters harvest brief voice samples from social media videos or recorded customer service calls, then feed them into AI models that can create entire conversations. The result is a flawless impersonation capable of tricking even the most cautious savers. Meanwhile, the rise of large language models enables scammers to craft personalised phishing emails devoid of the grammatical errors that once gave them away.
To fight back, the FCA is turning to AI itself. Its new ‘FraudNet’ system uses machine learning to identify suspicious patterns across millions of transactions in real-time. But as one cybersecurity insider put it, ‘This is an arms race where the bad guys also have AI. We’re not sure who will win.’ The agency has also formed a rapid-response unit with the National Crime Agency to track and dismantle fraud networks, though many operate from jurisdictions where extradition is unlikely.
For the average saver, the immediate advice remains cautious: never trust unsolicited calls or messages, even if they appear from your bank. Instead, hang up and call back using the number on your bank card. The FCA also recommends enabling two-factor authentication and using dedicated banking apps that offer biometric logins. However, as AI improves, even these safeguards may falter. ‘The ultimate solution,’ Rathi concedes, ‘is to redesign our entire authentication framework from the ground up, moving away from knowledge-based questions to behavioural and biometric markers that are harder to replicate.’
This emergency intervention signals a wider recognition that our financial system’s security perimeter has been breached by AI. For savers, the message is clear: trust nothing you hear or see without independent verification. For now, the FCA’s measures may provide a temporary shield, but the long-term battle against AI fraud has only just begun.
