Google researchers are fighting the company to tell the world what they know about AI
Current Google workers in Britain say post-employment contract language could deter them from discussing AI’s potential effects publicly.
TL;DR
- Google DeepMind workers in Britain are challenging non-disparagement language in their employment contracts, saying it could restrict public discussion of AI’s potential effects.
- Google says it does not enforce the clauses after staff leave and says it has removed the post-termination provision from terms for new hires.
- The company has not said it will revise existing contracts, leaving current employees’ legal position and the scope of protected public-interest speech unresolved.
Current Google workers in Britain told The Washington Post that non-disparagement terms in their employment contracts could prevent them from speaking publicly about AI’s possible impacts. Some Google DeepMind staff reviewed their contracts after a former colleague publicly criticized the company’s military AI work. [1]
One employee said the restriction could create uncertainty even if it would not ultimately be enforceable, because workers may avoid speaking rather than risk a legal or career dispute. The Post reported that dozens of employees discussed the clause internally and that some objected to its language. [1]
Google said it would not enforce the provision against former employees and said it had updated terms for new hires to remove the post-termination part of the clause. The report did not say that the company had changed existing contracts or detail what internal channels current staff can use to raise concerns. [1]
The Washington Post article drew a substantial Hacker News discussion, where commenters debated legal protection, workplace retaliation and the public interest in hearing from AI researchers. That discussion documents public reaction; it does not independently verify the contract terms or the employees’ account. [1] [2]
Why it matters
The dispute concerns whether researchers inside a major AI developer can speak about potential harms after leaving the company, and whether new-hire policy changes address concerns held by current staff. The reported gap between existing and updated contracts leaves the practical protection for current workers unsettled.
Editor's note
The contract concerns are attributed to current employees and company statements are presented separately. TODO-review: confirm legal terminology for UK non-disparagement clauses.