What happened
OpenAI removed the non-disparagement provision from its off-boarding paperwork after a former researcher declined the standard separation agreement and forfeited vested equity worth around $2 million, according to CryptoBriefing's report Tuesday. The clause reportedly tied continued ownership of vested units to a lifetime commitment not to criticize the company publicly. Under the original terms, breaking the clause could claw back equity that had already vested, an unusually aggressive posture even by Silicon Valley standards.
The researcher's decision to walk rather than sign forced the issue into the open. Within days, OpenAI told former employees the clause would not be enforced and that the language was being pulled from future agreements. CEO Sam Altman posted on X that he had not been aware of the provision, called it embarrassing, and said the company was working to make things right.
The company has not published the revised template.
Why it matters
The story matters beyond OpenAI's HR desk. Frontier AI labs sit at the center of the debate over safety, alignment, and disclosure, and gag clauses on ex-employees directly restrict the information flow that regulators, journalists, and researchers rely on. A clause that puts millions of dollars on the line for public criticism is not a standard NDA.
It is a permanent muzzle backed by financial penalty. The reversal is a rare case of a single individual's exit forcing structural change at a company valued in the tens of billions. It also lands with the SEC and other regulators paying closer attention to whistleblower protections in tech, where equity-linked silence agreements have drawn scrutiny in adjacent industries.