AI News
All stories
Official

OpenAI begins embedded assessments with Apollo Research, METR, Transluce and Goodfire

OpenAI says it has started the first phase of embedded assessments by outside evaluators of its model alignment, safeguards and monitoring.

Safety & Security · · 1 min read

OpenAI said it has begun the first phase of embedded assessments with four outside evaluators: Apollo Research, METR, Transluce and Goodfire. The post is dated November 13, 2026. According to the post, the groups will independently evaluate OpenAI's model alignment, its safeguards and its monitoring systems, which are the three areas the company names as the scope of the work.

The post is short. It does not say what "embedded" means in practice, how the outside groups will get access to the company's systems, how long the first phase will run, or whether any results will be published. Without those details, the announcement establishes who is involved and what they are assigned to examine, but not how deep the reviews go or what would count as a finding.

What changed

OpenAI began the first phase of embedded assessments by outside evaluators Apollo Research, METR, Transluce and Goodfire, covering its model alignment, safeguards and monitoring.

Type: Other

Why it matters

Independent testing of alignment and monitoring gives companies that buy OpenAI models a third-party view of its safeguards, though the announcement reports no findings yet.

What you can do

Ask model vendors which independent evaluators have reviewed their alignment, safeguards and monitoring, and what those reviews covered.

Original source

Further reporting

Where we saw it

3 sources

Forward Future (Matthew Berman), Latent Space, OpenAI

Apollo ResearchGoodfireMETROpenAITransluce

AI wrote this summary from the sources listed, and nobody reviewed it individually before publication. Check the original source before you rely on it. How it works