Anthropic CEO says humans hallucinate more than AI models

Anthropic CEO Dario Amodei has claimed that today's artificial intelligence (AI) models might hallucinate less frequently than humans.

Amodei made these remarks during a press event at the company’s first developer gathering, Code with Claude, held in San Francisco.

He addressed the broader concerns about hallucinations, instances where AI systems generate false or misleading information, and their impact on the pursuit of artificial general intelligence (AGI).

“It really depends on how you measure it, but I suspect that AI models probably hallucinate less than humans, though they do so in more surprising ways,” Amodei noted.

To note, Amodei is among the most optimistic voices in the AI industry regarding the timeline for AGI. In a widely shared paper published last year, he predicted AGI could emerge as early as 2026.

According to Amodei, hallucinations are not a major roadblock on the path to AGI.

“Everyone’s always looking for these hard blocks on what AI can do. They’re nowhere to be seen. There’s no such thing,” he stated.

That said, his view contrasts with that of other industry leaders as earlier this week, Google DeepMind CEO Demis Hassabis expressed concern that current AI systems still contain too many gaps, failing basic questions and undermining reliability.