That's the kind of situation I have in mind when thinking about how AI could harm humanity.
If AI would hurt humanity, that'd be thought escalation between nuclear countries. Either via false positive, like here, or by assisting voluntary attempts to setup false flag situations (typically via fake contents to stir unrest)
I'm most surprised that we actually heard about this and it wasn't buried, because it's a terrible look for the military and illustrates exactly what the AI alarmists are talking about lately, totally counter to what the administration and the DOD are pushing