Extinction riskFrontier labResearch scientist, lead of the Alignment Stress-Testing team, Anthropic1 entry on the record

X@EvanHub
Websitealignmentforum.org

Evan Hubinger

Research scientist who leads Anthropic’s alignment stress-testing team and, answering a colleague’s resignation, put his own estimate that AI kills all humans within a decade above ten percent, adding that the company has no plan yet for aligning superintelligence.

Roles

Oct 2024 – presentResearch scientist, lead of the Alignment Stress-Testing team, AnthropicStart date not sourced; the role is attested from this talk onward. Press on 2026-09-09 described him as Anthropic's alignment science lead.Source →

On the record

9 Sept 2026 · Post · X
Jacob is correct here—we really do earnestly believe AI could kill all humans! I personally think it is >10% within the next decade. I believe Anthropic is trying its best, but we do not yet have a plan to solve alignment for superintelligence and are not clearly on track to.
Research scientist, lead of the Alignment Stress-Testing team, Anthropic at the time
Replying to Jacob Coxon's resignation thread, as quoted by OfficeChai; the post is at x.com/EvanHub/status/2097497037956891126.
Anthropic Alignment Science Lead Evan Hubinger Says There's A More Than 10% Chance AI Could Kill All Humans Within Next Decade →