Anthropic & Pentagon: AI Ethics in Military Contracts

0:00I'm Corey, and this is your Pulse of the City, Durham News Today. Anthropic, the AI company known for its strict safety rules, is restarting talks with the Pentagon. The goal is to revive a contract worth up to $200 million. At the center are limits on using their clawed models for mass surveillance of Americans or fully autonomously the weapons. CEO Dario Amode wants to keep those bands in place while letting the military use the tech for other needs. Earlier this year, things fell apart during contract renewal talks. Anthropic had already put its AI on classified Pentagon networks for intelligence and cyberwork. But they refused broad access for any legal military mission, saying current AI isn't safe enough for life and death choices without human oversight. The standoff drew strong reactions. The Pentagon labeled Anthropic a supply chain risk blocking federal use of their models. Global liberties groups warned that vague contract language could harm privacy rights. Tech firms and advocates even urge leaders to drop the blacklist.

1:03Now, Amode is meeting again with Pentagon under Secretary Emil Michael. They're seeking a compromise before the risk label causes lasting damage. Anthropic insists on sticking to their ethical lines, no matter the cost. This could shape how AI companies deal with the military, one of the biggest buyers of advanced tech. A deal with real guardrails might encourage others to push for limits, influencing civilian tools and phones, banks, and schools down the line. D&N is sponsored by our partner. Check out the episode description. Your bedtime ritual just got better. Built in sound, built in comfort. S-O-L-I, solelypillow.com.