I imagine: AI vulnerability analysis is going to find something, it will be reported by a researcher and ignored as chaff, and then separately, later, someone

I imagine: AI vulnerability analysis is going to find something, it will be reported by a researcher and ignored as chaff, and then separately, later, someone will build it into an exploit and compromise a piece of critical national infrastructure