We don’t have to have unsupervised killer robots | The Verge

AIReportTechWe don’t have to have unsupervised killer robotsAI companies could stand together to draw red lines on military AI — why aren’t they?by Hayden FieldFeb 27, 2026, 4:18 PM UTCLinkShareGift Image: Cath Virginia / The VergePart OfAI vs. the Pentagon: killer robots, mass surveillance, and red linessee all updates Hayden Field is The Verge’s senior AI reporter. An AI beat reporter for more than five years, her work has also appeared in CNBC, MIT Technology Review, Wired UK, and other outlets.It’s the day of the Pentagon’s looming ultimatum for Anthropic: allow the US military unchecked access to its technology, including for mass surveillance and fully autonomous lethal weapons, or potentially be designated a “supply chain risk” and potentially lose hundreds of billions of dollars in contracts. Amid the intensifying public statements and threats, tech workers across the industry are looking at their own companies’ government and military contracts wondering what kind of future they’re helping to build.While the Department of Defense has spent weeks negotiating with Anthropic over removing its guardrails, including allowing the US military to use Anthropic’s AI kill targets with no human oversight, OpenAI and xAI had reportedly already agreed to such terms, although OpenAI is reportedly attempting to adopt the same red lines in the agreements as Anthropic. The overall situation has left employees at some companies with defense contracts feeling betrayed. “When I joine...