A series of incidents and stark new warnings from AI researchers has pushed the once-abstract risk of self-improving AI into the center of public debate. CFR’s Connor Martin unpacks what’s driving the sudden alarm and why the industry’s competing incentives could make a slowdown unlikely.