Employees from major AI labs are calling for coordinated government action on the pace of frontier AI development.
The Decoder reports that staff from leading companies want the US government to pursue international coordination, especially as automated research systems become more capable. Their argument is practical: one company or country cannot slow development alone if competitors keep accelerating.
The statement reflects growing concern that AI systems could help speed up AI research itself. If that feedback loop becomes powerful, ordinary corporate risk management may not be enough to control deployment pressure.
The proposal is not a detailed treaty or a binding pause. It is a public signal from people inside the industry that governance needs to move faster. The hard part is turning that concern into rules that are enforceable, internationally credible, and flexible enough to handle rapid technical change.