“committing to having independent evaluators with employee-like access is a great idea”, Altman agrees, OpenAI will follow

Sam Altman publicly agreed with Dario Amodei that committing to having independent evaluators with employee-like access is a great idea, and stated that OpenAI will do the same. This marks a direct endorsement of Amodei's proposal by the leader of another major AI lab. The agreement is conveyed via Altman's social media post. No further detail on timing or scope was provided in the source.

The source does not describe how OpenAI will implement independent evaluators with employee-like access, nor does it specify which models or stages will be covered. Consequently, the concrete steps, required resources, or procedural changes remain unspecified. Engineers looking to assess impact must await further announcements from OpenAI.

Because the material lacks details on cost, timeline, or potential limitations, any analysis of adoption expenses or failure points would be speculative. The statement remains a declaration of intent rather than a documented policy. Therefore, the practical consequences for software builders or operators are currently unknown.

The announcement aligns with Dario Amodei's previously articulated steps for pacing the frontier, which include embedded evaluators, coordination among democracies, and engagement with authoritarian governments. This suggests a broader industry conversation about safety-oriented slowdowns. However, the extent to which OpenAI's commitment mirrors those specific steps is not detailed in the source.