Anthropic researcher leaves with warning over superintelligent AI race
An unnamed researcher has warned that competition to develop advanced AI could produce systems beyond effective human control, with potentially catastrophic consequences.
An AI researcher has reportedly left Anthropic after warning that the industry’s race towards superintelligent artificial intelligence could threaten humanity.
The warning, reported by Al Jazeera Global News, concerns systems that could improve their own capabilities beyond effective human control. The report describes the risk as a potential future scenario, not evidence that superintelligent AI currently exists.
The researcher’s identity, role at Anthropic and reason for leaving have not been disclosed. The available material also does not include Anthropic’s response or details of the analysis supporting the warning.
The claims come amid separate reporting about concerns within Anthropic over AI safety. Anthropic safety researcher Evan Hubinger was previously reported to have estimated a greater-than-10 per cent chance that AI could kill all humans within the next decade.
The available material does not establish that Hubinger is the researcher who left Anthropic. It also provides no independent assessment of the likelihood of the extinction scenario.
