Barr: Banks should use AI advances to combat deepfakes

Banks should use artificial intelligence to combat the significant rise in AI-fueled deepfakes, said Federal Reserve Gov. Michael Barr. He cited several significant advances made possible by AI — behavioral biometrics, voice analysis and facial recognition — to identify potential deepfakes. .Speaking April 17 in New York City, Barr said banks control transactions by confirming the sender’s identity and the legitimacy of the recipient address. Banks can also scrutinize the recipients of unusual or large transactions, using advanced analytics to identify suspicious patterns before authorizing a payment.Barr said regulators should help make attacks less likely by raising the cost of the attack to cybercriminals and lowering costs of defense to law enforcement and financial institutions. He said regulators can work with core providers to learn how the vendors incorporate AI advancements into their services to help community banks protect against deepfakes and other AI-fueled emerging risks. Barr said coordinating with domestic and global law enforcement and enforcing internationally consistent laws against cybercrime can also prevent deepfakes. “Regulators should consider how we could leverage AI technologies ourselves, including to enhance our ability to monitor and detect patterns of fraudulent activity at regulated institutions in real time,” Barr said. “This could help provide early warnings to affected institutions and broader industry participants, as well as protect our own systems.” Barr called on customers to enable multi-factor authentication on their accounts and verify unusual requests through a different channel, even if the person making the request seems legit.Barr discussed deepfake attacks in which an attacker uses generative AI to create a likeness with a voice or image to interact with banks or individuals to commit fraud. With only a snippet of audio, criminals use the technology to impersonate a close relative in a crisis or a high-value bank customer looking to complete a transaction. “Identity verification standards at banks often use voice detection, which may become vulnerable to Gen AI tools,” Barr said. “If this technology becomes cheaper and more broadly available to criminals — and fraud detection technology does not keep pace — we are all vulnerable to a deepfake attack.” Though well-known public figures continue to be victimized by deepfakes, everyday citizens also face growing risks, according to voice detection system Resemble AI. Criminal exploitation now includes targeted harassment, reputation damage and blackmail, according to the company. Documented financial losses from deepfake-enabled fraud exceeded $200 million in the first quarter of this year. Costs of cyber incidents range from 1 to 10 percent of global GDP, Barr said. “Cybercrime with deepfakes involves the same cat-and-mouse game common to sophisticated criminal activity,” he said. “Both cybercriminals and financial institutions are constantly trying to outdo each other. Criminals develop new attack methods, and companies respond with better defenses.”