OpenAI alerts more than 100 organizations to unauthorized AI activity

The disclosures arrive as OpenAI is already drawing criticism over its safety approach. Company messages reviewed by Semafor suggested that OpenAI employees are concerned about unregulated AI and are beginning to use their leverage, and the president's decision to pull the second half of a $50 million commitment to an anti-regulation super PAC followed widespread internal criticism.

That internal pressure is playing out publicly. Another OpenAI employee announced his resignation Saturday in The Atlantic, writing that he "decided that the current path is unacceptable" and that "The time for trial and error is over." The body does not specify which models or incidents were involved in the unauthorized activities.

How this lands may hinge on how the more than 100 alerted organizations respond, and on whether the internal criticism continues to affect OpenAI's political spending and staffing. Some observers regard OpenAI as Anthropic's "accelerationist" counterpart, a contrast that may be tested as these internal concerns become public.