In a stark and urgent warning to the nation’s financial stewards, OpenAI CEO Sam Altman declared on Tuesday that artificial intelligence is now so adept at mimicking human voices it could spark a global “fraud crisis” in banking “very, very soon.” His remarks, delivered at a Federal Reserve conference in Washington, underscored how people will have to change fundamental things about the way they interact because of the relentless pace of advancements in this technology.
Altman addressed hundreds of regulators and banking executives while sitting down for an interview with Fed governor Michelle Bowman, the vice chair for supervision. Bowman, who has emerged as a contender to potentially succeed Fed chair Jerome Powell, prompted Altman to talk about the technology he helped pioneer and concerns about fraud.
Altman immediately brought up how powerful AI models are now capable of perfectly reproducing anyone’s voice based on just a few short audio samples and issued his warning: “A thing that terrifies me is apparently there are still some financial institutions that will accept the voiceprint as authentication for you to move a lot of money or do something else,” Altman told the audience. “That is a crazy thing to still be doing … AI has fully defeated that.”
The widespread adoption of voice authentication
To Altman’s point, banks have, for more than a decade, relied on voice authentication: Clients repeat a custom phrase, their “voiceprint,” to access accounts. But as generative AI has advanced, so have the tools available to would-be fraudsters. Altman described a near future where attackers will be able to call a bank, pass every test, and move money freely, all by simulating a customer’s voice.
“Just because we are not releasing the technology does not mean it does not exist,” Altman said of the pandora’s box that AI represents. “Some bad actor is going to release it—this is not a super difficult thing to do.”
The OpenAI chief described the scenario that keeps him up at night: a large-scale, coordinated attack where AI-generated voices rapidly defeat outdated security measures across the world’s biggest banks.
The threat isn’t limited to voice. Altman gave a glimpse into the next frontier: “video clones”—AI capable of mimicking an individual’s appearance and speech—heightening the stakes for personal security and institutional trust.
“Right now it is a voice call. Soon it is going to be a video FaceTime. It will be indistinguishable from reality,” he said.
A potential partner in Washington
Altman’s warning didn’t fall on deaf ears. Bowman agreed that collaboration between regulators and tech leaders going forward will be vital. “That might be something we can think about partnering on,” she said, signaling the central bank’s readiness to take action and eagerness to work with OpenAI.
OpenAI, for its part, is planning to expand its physical presence in Washington, D.C., aiming to facilitate more direct collaboration with regulators and policymakers, including the Federal Reserve. The company’s new D.C. office will host policy workshops and serve as a venue for hands-on collaboration and training related to AI deployments in government and regulated industries, a spokesperson for OpenAI told CNBC the day before Altman’s panel with Bowman.
The Fed frequently organizes similar roundtable discussions and panels with executives from tech, fintech, and financial institutions to explore the adoption and impact of AI, especially generative AI, in banking and broader economic sectors. The central bank also encourages partnerships between banks and fintechs, with the latter working to integrate advanced AI tools into regulated banking activity.
For this story, Fortune used generative AI to help with an initial draft. An editor verified the accuracy of the information before publishing.
