WatcherGuru
A researcher who worked at both OpenAI and Anthropic has resigned, citing fears that the accelerating AI race will produce systems capable of destroying humanity before adequate safeguards are in place. The departure is notable because the researcher worked at the two organizations most publicly associated with AI safety as a core mission.
The researcher's own words carry the weight here: "The people building AI earnestly believe that it could kill us all by the end of the decade." That framing is not a fringe position from an outside critic but an insider account of the working assumption inside the labs themselves.
Why it matters
When safety-focused researchers exit safety-focused labs over safety concerns, the signal is qualitatively different from external criticism. It suggests the competitive dynamics of the AI race are eroding the internal culture that was supposed to act as a check on deployment speed. Anthropic was founded explicitly by OpenAI alumni who believed the original lab was moving too fast; a researcher who passed through both and still found the environment untenable is a meaningful data point.
Market impact
Near-term, the resignation adds pressure to the regulatory conversation around frontier AI development. Policymakers in Washington and Brussels are already debating mandatory safety evaluations and compute thresholds; a high-profile insider defection citing existential risk strengthens the hand of those pushing for binding rules. For investors in AI infrastructure and the public cloud names exposed to frontier model compute, tightening regulation is the tail risk worth watching.
Frequently asked questions
-
Why did the researcher leave both OpenAI and Anthropic specifically?
The researcher cited fears that the competitive AI race is producing systems capable of destroying humanity before adequate safeguards exist, a concern they found unresolved at both labs despite their public safety mandates.
-
What did the researcher say the people inside the labs actually believe?
The researcher stated directly that the people building AI earnestly believe it could kill all of humanity by the end of the decade, framing this as an internal working assumption rather than an external fear.
-
Why does this resignation carry more weight than typical AI safety criticism?
The researcher worked inside the two organizations most publicly committed to AI safety as a founding principle, making this an insider account of cultural failure rather than an outside critique.
-
How could this resignation affect AI regulation in the US and Europe?
High-profile insider defections citing existential risk strengthen the hand of policymakers already debating mandatory safety evaluations and compute thresholds for frontier AI models in Washington and Brussels.
-
What is the market risk for investors in AI infrastructure if regulatory pressure increases?
If binding safety rules advance, the deployment timeline and cost structure for frontier AI shifts materially, with cloud providers and infrastructure names most exposed to frontier model compute facing the greatest impact.
CoinDesk
CryptoSlate
Crypto News
TheBlock
WuBlockchain