Washington Post says U.S. and Russia stripped safety clauses from AI weapons talks

The Washington Post reported that diplomats from the United States and Russia rewrote parts of a draft agreement on lethal autonomous weapons during negotiations in Switzerland earlier this month, removing several safeguards tied to the use of AI in warfare. The talks had brought together hundreds of diplomats in a large United Nations conference room and were aimed at producing what could have become the first agreement governing the use of lethal AI weapons. According to three people familiar with the negotiations, the final day shifted into a tense confrontation. Over roughly 15 hours, U.S. and Russian diplomats made intensive edits to the text. Those changes allegedly removed language requiring such systems to operate in a "predictable" and "reliable" manner, deleted a provision calling for ethical considerations in the use of AI weapons, and struck a clause that would have required human review of AI-generated military targets before a strike. The report said the narrower text showed how Washington and Moscow worked to weaken global rules for lethal AI weapons, as the Pentagon accelerates the integration of the technology into battlefield strategy.