An artificial intelligence researcher has quit the industry, accusing two leading US companies of "gambling with our lives" in the race to develop increasingly powerful AI systems.

The researcher, who previously left OpenAI to join Anthropic, cited concerns over the development of AI models capable of self-improvement beyond human control as the reason for their departure.

Anthropic safety executive Evan Hubinger responded publicly in support of the departing researcher, posting on X that the company genuinely holds serious concerns about the technology it is building.

"We really do earnestly believe AI could kill all humans!" Hubinger wrote, adding that he estimated the risk at more than 10 percent over the next decade.

The warning centers on the prospect of superintelligent AI systems that could improve themselves at a pace and scale beyond human oversight or intervention.