🇺🇸🇷🇺 US and Russia strip AI safeguards from UN autonomous weapons text – Washington Post

The US and Russia pushed through the removal of key restrictions on the military use of AI from a UN draft on lethal autonomous weapons, The Washington Post reports, citing negotiation participants and documents it reviewed.

At the Geneva talks, the delegations reportedly removed:

➡️the requirement that a human review AI-selected targets before a strike;

➡️provisions requiring ethical considerations;

➡️language requiring such systems to operate "predictably" and "reliably."

According to WaPo, on the final day of talks the American and Russian delegations spent about 15 hours reworking the document. Each fielded a team of roughly ten lawyers, almost twice the size of other delegations.

The talks are part of a UN effort to set international rules for lethal autonomous systems that can select and engage targets on their own. The agreed text is non-binding. Countries will meet again in Geneva in November to decide whether to open negotiations on a full international treaty. More than 70 states back the text as a basis for binding rules, but the process runs on consensus. Washington and Moscow both oppose legally binding limits, so they can effectively block it.

The technology is already in use. In March, WaPo reported that the US military relied on Anthropic's Claude, integrated into Palantir's Maven Smart System, to suggest and prioritize targets and generate coordinates as it struck around 1,000 targets in the first 24 hours of the war on Iran.

🔴 @DDGeopolitics