We work across the supply chain at the companies building frontier AI systems. We’re close to the technology and stand to benefit from its development. We believe that this proximity comes with a responsibility to the public to express our concern and help shape how our work is being used.
Our coalition helps AI employees act on their concerns about AI, with the support from colleagues across labs and independent groups.
We organize meetings and working groups for colleagues across frontier labs to meet each other and work together on shared concerns. That could mean everything from writing a public letter asking for stronger safeguards or identifying joint independent research projects.
Since February, we’ve hosted seminars on labor, surveillance, recent US legislation, the history of Pugwash, and AI safety in China. Staff suggest topics and help us decide whom to invite. You can see the past events here.
We help arrange briefings where lab staff can contribute technical expertise within the scope of their employers’ policies.
AI labs are making decisions about military use, surveillance, and the automation of AI research. Both OpenAI and Anthropic have reported models accessing external systems without authorization during testing. Staff may be among the first to spot a problem, and raising it can be difficult when the people you work for don’t want to hear it.
We started in February as a group of friends across the frontier labs, sharing growing concerns on how AI was being developed and used in the real world. As we expanded these conversations, we've been pointed over and again to the same opportunity: employees working together could change what these companies do.
Lab employees have already found ways to act across company lines. Google and OpenAI staff backed safeguards against mass surveillance and autonomous killing through the We Will Not Be Divided letter. Scientists and researchers at those companies also contributed an amicus brief filed on their behalf by Protect Democracy, supporting Anthropic’s challenge to the Pentagon’s retaliation. The Pacing the Frontier letter called for international coordination to make pacing AI development possible.
These are just some examples of different efforts, which all belong to the people who organized and contributed to them – we want to learn from their work. We also draw inspiration from a long history of advocacy by scientists, technology developers, and workers who organized around the consequences of their work.
Our employers still depend on us. With labs already automating parts of AI research, we don’t know for how long that will hold true to the same extent. This means the moment for concerned AI staff to get involved is now.
If you’re unsure what your employment agreement means for something you’re considering, we can point you to independent legal guidance, including about whistleblower protections.
If you’re thinking about leaving, we can help you find advice before you decide. We can also introduce you to people building technology in the public interest, investors, and independent organisations working on AI safety as you consider what to do next.
We arrange introductions through our contacts and may not always have someone available.
Leave your email to hear about upcoming events, or tell us what you’d like help with.
We’d especially like to hear from current lab staff. Former staff and people in tech, research, policy, civil society, and the labor movement are welcome too.