A us presidential candidate, before the elections next November (poling at least 1% at the time) says something for public consumption that's quoted, filmed, taped, or otherwise publicly knowable afterwards to the effect of them having an intention to bring their authority to bear if they can to address some kind of near term AI risk.

Naive understandings particular risks or of how difficult it might be to take action do not disqualify. For example the following would resolve true: "my administration would be prepared to act immediately if we got wind of one of those things being involved in starting a war" or "if those things start messing with the stock market, I think we'll have to step in right away", or "any word of those things telling people how to make new viruses and we're going in and pulling the plug", etc.

Vague language including other branches of government or discouraging without a claim about decisive use of presidential power will not resolve positive, such as "we should have a referendum if it's going to affect us", "it's moving too fast", "we're going to have to regulate it", "more safety should be required", "congress needs to act", etc.