Research shows that AI models can penetrate computers and copy themselves

Scientists who observed AI chatbots copying themselves and launching hacking attacks in an experiment say: “Autonomous AI self-replication is no longer hypothetical.”