Overview
- A UN meeting under the Convention on Certain Conventional Weapons will open in Geneva on November 16 to discuss restrictions or bans on lethal autonomous weapon systems.
- Experts define those systems as AI-enabled weapons that can identify, select and engage targets without human intervention during those tasks.
- States disagree sharply on acceptable human control, with the United States calling for "appropriate levels of human judgement" and France insisting on a human decision loop while permitting partially autonomous systems.
- UN and humanitarian leaders warned this year that autonomous targeting risks crossing a moral red line, and some AI leaders have said current frontier systems are not reliable enough to run fully autonomous lethal weapons.
- Delegates will weigh short-term risks such as limited battlefield deployments and drone swarms against long-term effects including a faster arms race, legal gaps and new civilian harm from automated targeting decisions.