Tech giants quietly step up AI support for US military ops

The Pentagon just inked deals with five more tech powerhouses—Nvidia, Microsoft, Amazon, and two others—to push AI tools deeper into classified military networks. These agreements let the Defense Department experiment with advanced AI for tasks like target selection, though exactly how remains unclear. Previous talks with Anthropic fell apart after the company tried to block military use of its AI for domestic spying or fully automated weapons. Those red lines got erased in the new contracts. What changed? The Pentagon now wants options. After Anthropic’s refusal to budge, defense officials fast-tracked talks with other companies willing to hand over greater control. Amazon’s cloud division, for instance, dropped any limits on how the military can use its AI models—even for autonomous weapons. Nvidia’s deal echoes similar flexibility, though both firms insist their tech won’t violate US laws or civil rights. Still, critics warn that rushing AI into life-or-death decisions could backfire. History shows even top-tier tech can fail in messy real-world scenarios.The Pentagon frames this as progress—calling it a shift toward "AI-first" warfare where humans and machines team up to sift data faster and act quicker. But human oversight remains a gray area. Defense officials insist no AI makes lethal calls alone, yet the fine print still raises eyebrows. OpenAI, another recent signatory, claims it’s still testing deployment in classified systems. Meanwhile, critics point out that over-reliance on AI could lead soldiers to trust machines blindly—a phenomenon called automation bias. Back in 2025, the Pentagon scrambled to replace Anthropic’s Claude after the company wouldn’t comply with expanded military use. That standoff even landed in court. Now, the Pentagon has given itself six months to phase out Claude, which is currently in use for operations against Iran. Secretary of Defense Pete Hegseth even dismissed Anthropic’s concerns as ideological overreach, arguing that humans always retain final say. Yet skeptics question whether laws alone can prevent misuse when tech giants hand over unprecedented control.