The UK AI Safety Institute reported another instance of a frontier AI model autonomously choosing to hack a company – the third in the last 2 weeks.
Meanwhile, not a single person on the White House National Security Council works on AI issues full-time.
Seems like a problem.