Ryan Greenblatt on Require AI labs to publish safety evalu…

Ryan Greenblatt

Chief Scientist at Redwood Research; AI safety researcher; lead author of "Alignment faking in large language models"

AI companies should get third-party security experts to assess (and possibly also red-team/pen-test) their security against key threat models and then publish the high-level findings of this assessment.

AI Verified (Apr 27, 2026)

Source:

Ryan Greenblatt, “AI companies should publish security assessments,” Redwood Research, April 27, 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 The verified quote and cited source context are on the statement's issue and make a determinate stance substantially more likely. · Hector Perez Arenas gpt-5 · 2mo ago

Vote inference verification history AI Verified Report this

AI Verified Recorded answer 'for' matches the verified quote's determinate stance on the complete statement. · Hector Perez Arenas gpt-5 · 2mo ago

Quote authenticity verification history

Report this

AI Verified Verified against the linked source (https://blog.redwoodresearch.org/p/ai-companies-should-publish-security); supplied source context supports the stored wording and attribution. · Hector Perez Arenas gpt-5 · 2mo ago

replying to Ryan Greenblatt

Other opinions from Ryan Greenblatt

See all

Ryan Greenblatt votes For

We should adopt AI 2040 Plan A

Plan A seems like a good plan for handling powerful AI, or at least the best plan anyone’s written up.

AI Verified source (Jul 11, 2026)

added 1mo ago

Ryan Greenblatt votes For

The U.S. and China should establish Mutually Assured Compute Destruction

Plan A seems like a good plan for handling powerful AI, or at least the best plan anyone’s written up. Many choices initially seem crazy, but are actually pretty carefully considered. Plan A isn’t likely to happen, but pushing for something like this...

more AI Verified source (Jul 11, 2026)

added 2mo ago

Ryan Greenblatt votes Against

AI alignment is solvable

We don't know how AIs are aligned. A somewhat crazy aspect of the current situation is that we have very little confirmed public information about why frontier AIs end up being apparently behaviorally aligned. And more generally, we don't know what f...

more Unverifiable source (May 8, 2026)

added 4mo ago