Sign or Lose Everything: OpenAI's Draconian NDAs
Key Points
- Former OpenAI employees revealed harsh NDA and equity clawback provisions
- Departing staff faced losing vested equity if they criticized OpenAI publicly
- Sam Altman apologized and promised to fix the equity clawback clause
- Raised concerns about suppression of safety whistleblowers in AI labs
- Congress members called for investigation into AI industry NDA practices
In May 2024, reporting revealed that OpenAI's departure agreements included non-disparagement clauses that could claw back vested equity from former employees who spoke critically. After public backlash, Altman apologized and promised to remove the provisions.
OpenAI made employees sign agreements saying they'd lose their shares if they said bad things about the company. After people found out, Altman said sorry and they changed it.
Sides
Critics
No critics identified
Defenders
Apologized and committed to removing clawback provisions
Noise Level
Forecast
Legislative action on AI whistleblower protections is likely. Other AI labs will preemptively review their own equity and NDA terms.
Based on current signals. Events may develop differently.
Timeline
OpenAI removes clawback provisions
Company announces it will release former employees from non-disparagement obligations
Former employees speak out, Altman apologizes
Altman calls agreements a mistake and says he was unaware of the specific provisions
Vox reports OpenAI equity agreements include non-disparagement clawback
Departing employees could lose millions in vested equity for criticizing the company