Sam Altman makes big revelation, admits GPT-5 makes him nervous

OpenAI CEO Sam Altman has sparked fresh debate with his candid remarks about the upcoming AI model GPT-5.

Speaking on the This Past Weekend podcast with Theo Von, Altman described testing GPT-5 in a tone that felt more like a suspense thriller than a product launch.

He said the model “feels very fast,” and admitted feeling genuinely uneasy during some testing sessions.

Despite leading its development, Altman said GPT-5 reminded him of the Manhattan Project, the US initiative that built the first atomic bomb.

Altman didn’t share any technical details but made it clear the model’s behaviour left him deeply unsettled.

He also criticised the current state of AI regulation, stating that “there are no adults in the room,” and that oversight has fallen far behind the pace of development.

While he might have meant to highlight potential risks, his comments painted a troubling picture, suggesting OpenAI might not fully understand the power of what it has created.

For those unversed, GPT-5 is rumoured to bring major improvements over its predecessor, GPT-4, possibly pushing artificial general intelligence (AGI) closer to reality.

That said, Altman’s dramatic tone has drawn criticism, with some calling it exaggerated and unhelpful.