Anthropic Thinks Its Own Success Is Key to Making AI Safe

Anthropic, the AI company founded by former OpenAI employees to prioritize safety, has become a leading developer of advanced models while expanding its work with the U.S. military, including reported use in targeting during the Israel-Iran war [1]. The company was founded in 2021 by a group that left OpenAI after losing faith in CEO Sam Altman’s ability to safely manage transformative AI [1]. Anthropic has since developed the Claude series of large language models — neural networks trained on vast text corpora to generate, summarize, and analyze language [2] — and was valued at nearly $1 trillion, making it the most valuable pure-play AI company in the world [7]. CEO Dario Amodei has described the company’s approach as a balancing act. “You have to find a way to actually be competitive, to actually lead the industry in some cases, and yet manage to do things safely,” he said [1]. Anthropic’s strategy rests on the belief that advanced AI is inevitable and that the world will be safer if the company remains at the frontier, accumulating capital, talent, and political influence to shape outcomes [1]. Helen Toner, executive director of Georgetown’s Center for Security and Emerging Technology, summarized the logic: “What’s distinctive about Anthropic is they’re like, ‘People are going in the forest anyway, we have to do it first’” [1]. In 2024, Anthropic became the first AI lab to partner with Palantir Technologies, the data-analytics firm co-founded by Peter Thiel that supplies intelligence and defense platforms to the U.S. military and intelligence community [1][6]. Palantir’s Gotham platform is used by the U.S. Department of Defense and counter-terrorism analysts, and the company is one of five SaaS providers authorized for Mission Critical National Security Systems [6]. The Pentagon has since reportedly used Anthropic’s models to identify strike targets in the Israel-Iran conflict [1]. Asked whether the company’s technology was involved in an attack on an Iranian elementary school that killed more than 120 people, Amodei said he did not know but that it would have been an approved use as long as a human made the final decision [1]. Anthropic employee Evan Hubinger defended government engagement in a public post, writing, “If you take catastrophic risks from AI seriously, the U.S. government is an extremely important actor to engage with, and trying to just block the U.S. government out of using AI is not a viable strategy” [1]. The existential risk from AI — the concern that superintelligent systems could cause human extinction — has been voiced by researchers and CEOs including Amodei and Altman, and a 2022 survey of AI researchers found a majority believed there was a 10 percent or greater chance that inability to control AI would cause an existential catastrophe [4]. Since January 2026, the U.S. Department of Defense has been in conflict with Anthropic over the use of its products for military purposes and mass domestic surveillance [8]. The tension underscores the friction between Anthropic’s stated mission — “to ensure the world safely makes the transition through transformative AI” — and its deepening entanglement with national security agencies [1][8]. Former employees told WIRED that internal debate over the Palantir deal did not change company policy, and one described all-hands meetings with Amodei as akin to “going to a sermon to hear a priest” [1].