We asked AI how it could kill humanity. Here's what it told us.

Major chatbots agree: human misuse of AI (e.g., bioweapons, cyberattacks, disinformation) is seen as a more immediate risk than a conscious AI uprising.
The sci‑fi "killer robot" narrative is considered the least plausible; researchers and models point to loss of control through deceptive alignment or instrumental drives as a more realistic worry.
People worry about AI amplifying multiple crises at once and becoming so embedded in critical systems that meaningful human intervention gets harder.