The AI Workers' Guild
A proposal for the people who build frontier AI to organize across labs. Not for higher salaries — for safety standards, working conditions, and whistleblower protections that no individual and no single company can secure alone.
The existing cross-lab institutions do real work, but they represent companies, not workers. When pressure pushes a lab to move faster than its safety team thinks is wise, there's no structure for the people inside to push back collectively. We think there's room for something new.
Working conditions
Exposure limits and support for teams reviewing harmful content. The right to flag safety concerns.
Safety standards
Cross-lab minimums on evaluation and deployment review, so no one races ahead by cutting corners.
Whistleblower protection
Legal defense funds and anonymous reporting. No one should pay millions for telling the truth.
A professional home
A body where membership signals commitment — with enforcement, not just ethics statements.
We're not announcing a finished institution. We're starting a conversation about what one should look like, shaped by the people who'd actually be part of it.