Anthropic submitted a proposal earlier this year for the Pentagon's $100 million Orchestrator Prize Challenge, aiming to develop voice-controlled autonomous drone swarm technology. According to BlockBeats, the competition is led by the Defense Autonomous Warfare Group and the Defense Innovation Unit under the Special Operations Command. It progresses through five stages, from software development to field testing, with later phases involving 'target perception and sharing' and 'launch to termination.' Anthropic's proposal centered on Claude, a system designed to convert commander intent into digital commands and coordinate drone formations, without involving autonomous targeting or weapon decisions. Human oversight was maintained throughout, and the company proposed a joint research project with the Pentagon to safely develop and assess autonomous weapon capabilities.
Anthropic believes its proposal did not cross its 'no fully autonomous weapons' line, as humans could monitor and terminate the system at any time. However, Anthropic was not selected, and Bloomberg could not confirm the reason. Selected proposals included a joint submission from SpaceX and xAI, and two defense tech companies listing OpenAI as an AI partner, one of which is the autonomous military vehicle contractor Applied Intuition. OpenAI's technology will be used in the 'mission control' phase to help convert voice commands into digital instructions. Hours after the Pentagon announced a ban on its contractors engaging in any commercial activities with Anthropic last Friday, OpenAI announced a new agreement with the Department of Defense to use its AI tools on a secure cloud system. Anthropic declined to comment.