Alexander Meinke on Require AI labs to publish safety evalu…

Alexander Meinke

Head of research at Apollo Research, an AI-safety organization focused on evaluating advanced AI systems.

If you ask the model developers, Was the AI plotting to take over the world during training?, you want the answer to be a resounding no. The actual answer is: I don't know. Nobody checked.

AI Verified (Aug 12, 2026)

Source:

Alexander Meinke, quoted in Matteo Wong, “It May Be Time to Panic About AI,” The Atlantic, August 12, 2026.

2mo ago

Policy proposals and claims

Require AI labs to publish safety evaluations before deploying frontier models

votes For

Statement relation verification history AI Verified Report this

AI Verified Meinke says developers did not check whether models were plotting takeover during training, directly supporting the need for pre-deployment frontier-model safety evaluations. · Hector Perez Arenas gpt-5.6 · 2mo ago

Vote inference verification history AI Verified Report this

AI Verified Correct answer is for: Meinke's warning that developers did not perform this safety check strongly supports requiring labs to publish such evaluations before deployment. · Hector Perez Arenas gpt-5.6 · 2mo ago

Quote authenticity verification history

Report this

AI Verified The Atlantic (Aug. 12, 2026) directly attributes the stored wording to Apollo Research head Alexander Meinke; wording and date match the article. · Hector Perez Arenas gpt-5.6 · 2mo ago

replying to Alexander Meinke

Other opinions from Alexander Meinke

See all

Alexander Meinke votes For

AI labs should give independent evaluators employee-like access

AI companies should be able to answer some very basic questions about their training process, such as: Did the AI ever actively try to undermine its own alignment training while it was going through the training? The answer to this should be an unequ...

more AI Verified source (Sep 16, 2026)