Futurology

AI's hallucination problem still hasn't been solved, which means the rational choice is to second guess everything it says. In which case, is all the extra work generated worth the time?

AI is vulnerable to manipulation via data poisoning, and as Big Tech is so opposed to safety regulation, we've no idea how much this is happening. It would be in some countries' interest to spark a major security confrontation between China and America. That may not be what happened here, but there are surely people thinking and acting along these lines.

US military had close call after using AI for false intelligence report, sources say