Killer robots become battlefield norm as UN moves to draw AI weapons 'red lines'

As AI-powered autonomous lethal weapons become a fixture on battlefields around the world, the UN is set to open formal discussions on regulating them.

According to AFP, the UN will convene a meeting Monday at its headquarters in Geneva, Switzerland, to discuss measures to regulate or ban so-called "killer robots" — autonomous weapons systems capable of operating without human intervention.

Such weapons can identify, select and engage targets without any human input.

Anti-personnel landmines, long used in warfare, represent an early form of autonomous lethal weapon, but rapid advances in technology have dramatically expanded the scope of such systems. AI-equipped drones are the most prominent example.

These weapons can select and strike targets not predetermined by humans, and are capable of adapting to battlefield conditions without human involvement.

The use of autonomous lethal weapons has become routine in combat. Jean-Marc Rickli, head of global and emerging risks at the Geneva Centre for Security Policy, said the United States struck up to 1,000 targets a day inside Iran during the war that began in late February.

China's Defense Ministry described swarms of autonomous drones in January as "the new face of war."

As warfare involving autonomous lethal weapons intensifies, regulators are grappling with how much human control should be required in weapons deployment decisions.

The United States emphasizes "an appropriate level of human judgment."

France insists that humans must be involved in decision-making, and has banned fully autonomous lethal weapons while permitting only semi-autonomous systems.

UN Secretary-General António Guterres and Mirjana Spoljaric, president of the International Committee of the Red Cross, warned in a joint statement in August that the world risks "crossing a moral red line" by allowing machines to autonomously designate humans as targets.

Dario Amodei, CEO of AI company Anthropic, also argued this year that "cutting-edge AI systems are not trustworthy enough to power fully autonomous weapons."

Experts predicted that even if the international community discusses regulatory frameworks for autonomous lethal weapons at the meeting, the talks will address only a fraction of the problem.

Rickli said the meeting "does not cover all the issues surrounding AI on the battlefield."

Not only is the proliferation of autonomous AI weapons outpacing the discussions, but key capabilities — such as AI-driven target recommendations — are being left out of the agenda, limiting the prospects for meaningful solutions.

Gerald Mako, a researcher at the University of Cambridge, wrote in the Small World Journal that "the first country to normalize fully autonomous lethal strikes in an era of intense military competition will reset the expectations of every other country."