U.N. talks aim to rein in ‘killer robots’ as fears grow over autonomous weapons: NPR


An armed robot developed by Turkish defense companies is demonstrated at a facility in Duzce, Turkey, April 29.

An armed robot developed by Turkish defense companies is demonstrated at a facility in Duzce, Turkey, April 29.

Omer Urer/Anadolu via Getty Images


hide caption

Omer Urer/Anadolu via Getty Images

GENEVA — Negotiations on possible rules for the use of deadly autonomous weapons, also known as “killer robots,” are underway at the United Nations in Geneva.

Proponents of autonomous weapons control say the U.N. Conference on Certain Conventional Weapons (or CCW), which meets this week once every five years, is the last chance to reach agreement on rules before it is too late.

Warfare technology has advanced rapidly in recent years, with some experts warning that the world will be just months away from having weapons that can autonomously choose targets and make decisions to kill.

In Ukraine and Gaza, AI-assisted weapons have already been used to eliminate targets, with Ukrainian civilians reporting being “stalked” by drones while shopping or going to school. But these are still controlled remotely by humans.

Humanitarian organizations say the decision to kill – and responsibility for that choice – must remain in human hands. Ahead of this week’s meeting, the United Nations, the International Committee of the Red Cross and the Pope have all called for legally binding rules. UN Secretary-General Antonio Guterres has warned that allowing fully autonomous weapons to kill would be “crossing a moral red line”.

What are the risks?

The Red Cross is particularly concerned that the existing rules of war – the Geneva Conventions – cannot be fully respected by autonomous weapons. War is often chaotic and confusing, and the situation can change quickly. Under the Conventions, wounded or surrendering soldiers can no longer be targeted.

Georgia Haines, legal counsel at the Red Cross, worries that autonomous weapons will not be able to distinguish between a soldier who continues to fight, a wounded soldier and a soldier who wants to surrender.

“The risk is that, based on the technology we’re seeing, they won’t be able to understand contextual cues in the same way a human can.” The risks of misidentifying a target are therefore incredibly high.

Gn headline

Logo