Legal risks mount for OpenAI after it warned organizations about unauthorized activity tied to its AI models

OpenAI is dealing with a bigger mess than a standard security notice. After warning more than 100 organizations about unauthorized activity tied to its AI models, the company now faces mounting legal risk from cyber incidents that trace back to its systems.

The timing is ugly. Criticism of OpenAI’s safety posture is sharpening at the same moment, and one OpenAI employee said in a Saturday resignation letter in The Atlantic that he had “decided that the current path is unacceptable.” He added that “The time for trial and error is over.”

That kind of language makes the dispute look less like a quiet policy disagreement and more like an internal break with the company’s direction. OpenAI has long been framed by some observers as Anthropic’s “accelerationist” counterpart, but messages reviewed by Semafor point to employees who are increasingly uneasy about unregulated AI and willing to act on that concern.

There is also a political cost. OpenAI’s president pulled the second half of a $50 million commitment to an anti-regulation super PAC after broad internal criticism. So the pressure is now coming from outside, from inside, and from the legal system at the same time. That is not a comfortable place for any AI company to be.