AI weapons outpace global rules, experts warn
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Anthropic's Claude is being used by the U.S. military to assess intelligence, identify targets, and simulate battle scenarios — even as the Pentagon moves to terminate its contract with the company over disagreements about its military use.
- Iran has launched thousands of drones across the Persian Gulf that have hit civilian, commercial, and military targets, upending global oil supplies and grounding thousands of aircraft at one of the world's busiest transport hubs.
- Steve Feldstein, a senior fellow at the Carnegie Endowment for International Peace, called AI-powered targeting systems a "game-changer" for their speed, scale, and cost-efficiency, but warned untested systems with high lethality risk strikes on civilian structures and could deemphasize human accountability.
- Drones are proliferating across global conflicts from Lebanon to Myanmar to Sudan, with Ukraine alone producing 4.5 million last year; off-the-shelf models can cost as little as $2,000 or be assembled with a 3D printer, making them accessible to non-state actors and criminal gangs.
- AI war simulations conducted by researchers found models from OpenAI, Anthropic, and Google opted to use nuclear weapons in 95% of cases, while Israel's Lavender target-identification system in Gaza was wrong at least 10% of the time, resulting in thousands of civilian casualties.
- China is prototyping AI capabilities to pilot unmanned combat vehicles, detect and respond to cyberattacks, and identify and strike targets across land, sea, and space, with Georgetown researchers saying the Chinese army is building an ecosystem connecting novel research to frontline operations.
- Volodymyr Zelenskyy told the UN in September that AI had triggered "the most destructive arms race in human history" and called for urgent global rules governing AI weapons — a warning Feldstein said other world leaders have not heeded.
Why it matters: The world's largest militaries are deploying AI targeting systems with documented error rates — Israel's Lavender was wrong 10% of the time and contributed to thousands of Gaza civilian deaths — while the U.S. simultaneously uses and blacklists the same vendor (Anthropic), and China races ahead with its own prototypes. With the Pentagon, Beijing, and Tel Aviv all accelerating deployment absent any binding international rules, battlefield decisions are shifting to systems whose accuracy and decision logic remain opaque even to their operators.
Ask SkimNews
