Sam Altman, the CEO of OpenAI, is widely recognized as the driving force behind the explosive rise of ChatGPT. However, he frequently raises alarms about the potential dangers that come with the advancement of artificial intelligence (AI).

Recently, the 40-year-old executive warned that the world may be on the brink of a major fraud crisis, as AI technology could allow bad actors to impersonate others with ease.

“What really scares me is that some financial institutions still accept voice authentication for authorizing large fund transfers or sensitive transactions. You simply say a phrase, and they proceed with it,” Altman said.

This is especially concerning because current AI capabilities allow for highly accurate voice cloning. “That’s a dangerously reckless practice. AI has essentially outsmarted most existing authentication methods, aside from passwords,” he elaborated, as reported by detikINET citing an interview with CNN.

His remarks were part of a broader discussion on the societal and economic impacts of AI, delivered during an event hosted by the Federal Reserve in Washington. While Altman warned of the risks, OpenAI has also urged the Trump administration to avoid implementing regulations that might restrict U.S. tech companies from keeping up with international competition in AI innovation.

Altman is not alone in expressing concern over AI-fueled fraud. The FBI has issued alerts about scams involving AI-generated voice and video cloning. Many parents have reported being deceived by fraudsters using synthetic voice technology to mimic their children's voices and fabricate emergency situations.

Earlier this month, U.S. officials revealed that someone had used AI to replicate the voice of Secretary of State Marco Rubio, making fraudulent calls to other U.S. officials including governors, foreign ministers, and members of Congress.

“I’m extremely worried that we’re heading toward a serious fraud crisis—and it’s coming soon. Right now, it’s voice. But very soon, it will be video and FaceTime calls that are indistinguishable from reality,” Altman warned.

He also revealed what keeps him awake at night: the thought of malicious actors misusing superintelligent AI before humanity has built sufficient defenses against such threats. For instance, adversaries might deploy AI to target the power grid or develop biological weapons.

Altman further expressed concern that humans might lose control over superintelligent AI systems—or worse, delegate too much decision-making power to them. Various tech companies, including OpenAI, are actively pursuing superintelligent AI development. According to Altman, the 2030s could usher in an era where AI surpasses human intelligence by a wide margin.