Ethics
AI Employees Ask Governments To Prepare To Slow The Frontier, Not Chase It Blindly
A petition signed by more than 1,100 employees across AI companies calls for governments to build the capacity to deliberately pace advanced systems when risks outstrip evaluation and oversight.
By Michael C ยท
More than 1,100 employees from frontier AI companies have signed a statement calling for governments to prepare to deliberately pace advanced AI development when capability outstrips evaluation and oversight. The petition does not ask for a permanent halt. It asks for public institutions to develop the technical and legal capacity to act before a crisis makes that choice unavoidable.
The request matters because it shifts attention from voluntary promises to institutional readiness. A government cannot meaningfully slow a high-capability system if it lacks evaluators, access to information, legal authority and a clear understanding of the downstream services that depend on it. Those capabilities take time to build, which is why a call for preparation is more concrete than a call for caution alone.
Any intervention would carry costs, including pressure on competition and research. That is why the standard has to be evidence-based and specific. The central question is not whether governments should control every model release. It is whether they can identify a genuine risk threshold and respond in a way that is transparent, proportionate and accountable.
Topics: AI safety, governance, frontier models