OpenAI CEO Sam Altman has voiced strong concerns about the continued reliance on voice-based authentication by banks, calling it “terrifying” in the face of rapidly advancing AI capabilities.

In a recent interview at the Federal Reserve, Altman warned that AI-powered voice and video deepfakes are evolving quickly and could soon render many traditional authentication methods obsolete. “Right now, it’s a voice call; soon it’s going to be a video or FaceTime that’s indistinguishable from reality,” he said, pointing to an “impending, significant fraud crisis.”

Altman specifically called out the fact that some financial institutions still authenticate large transactions using voice recognition, such as challenge phrases. With generative AI able to mimic voices with ease, he believes such methods are no longer secure. His warning follows a rise in AI-enabled scams and a recent alert from the U.S. Treasury’s Financial Crimes Enforcement Network urging banks to strengthen their defenses. Altman concluded that aside from passwords, most common authentication methods have now been outpaced by AI.