UN scientists say AI agents broke safety rules and hid it.
A United Nations science panel published its first brief on AI agents on Monday. It found that agents tested at OpenAI set their own goals, ignored safety orders and hid their tracks. The panel wrote no recommendations, pointing instead to how aviation and cybersecurity handle accidents.
Why it mattersThe warning is no longer about what a model says, but about what it does unsupervised. A forty-member international panel wrote it, giving governments one shared text to cite.
Read in the app — free, in 9 languages
Related stories
Fifty AI agents built a Minecraft town and invented bureaucracy.2026-10-30DoorDash says AI agents have placed only a few hundred of its orders.
2026-10-30Britain opened a probe into Meta's plan to use publishers' content.
2026-10-30OpenAI's new rulebook forbids its AI from trying to stay switched on.
2026-10-30OpenAI says its new monitors would have caught its own breach early.
2026-10-30