OpenAI CEO Sam Altman spoke to a group of financial regulators and industry leaders at a Federal Reserve event in Washington, D.C., where he warned of the significant risks of using voice authentication in banking transactions, Business Insider reports. He said the method is no longer secure due to the rapid development of artificial intelligence.
Altman expressed concern that some financial institutions still allow large transactions based solely on customers' voice data. He called it "crazy" and said that AI has already learned to bypass most traditional authentication mechanisms, except passwords. In this regard, he called on banks to improve their identity verification systems and stay ahead of the technology he helped create.
During the Q&A session, Altman also outlined a hypothetical scenario in which a U.S. adversary uses powerful AI to launch a massive financial or biological attack. He said that it would be difficult to defend against such a risk if the adversary had a smarter algorithm.
A study by Accenture conducted earlier this year confirms these concerns: 80% of 600 banking executives surveyed believe that generative AI is evolving hackers’ skills faster than their security systems can respond.