Will resolve positive if there is some kind of agreement that:

• Substantially concerns the US and China (and potentially other countries as well)

• Concerns AI directly (such as by regulating compute)

• In the process of being created, somehow used AI in a mediating role in the negotiation process. I will somewhat handwavily operationalise this as: "Where both parties pass information to the same AI model, in a way where that information gets used by the model in the same inference run, tree rollout, or other similar closely related computational step".

This last clause is a broad definition and can include things like: an AI chatbot being an actual mediator, some form of computational social choice model being involved, AI taking in both sides preferences and then being used to generate candidate agreements which humans then negotiate over, etc.

The question will not resolve true for things like bureaucrats using ChatGPT as a work tool to ask questions, or to proofread drafts of their legislation, or if either side privately uses their own AIs to work but there is no AI that both parties end up using.