Why It Matters
Congress has not enacted legislation specifically governing the Department of Defense's (DOD) use of frontier artificial intelligence models or their reliability in autonomous weapons systems. A Congressional Research Service (CRS) report updated Oct. 5 identifies potential congressional responses amid the Pentagon's dispute with Anthropic.
On Feb. 27, Reuters reported that President Donald Trump directed federal agencies to stop using Anthropic's technology. Defense Secretary Pete Hegseth subsequently directed DOD to exclude Anthropic from its supply chain, allowing up to six months to transition away from its products.
On Sept. 25, a divided D.C. Circuit panel upheld DOD's exclusion finding sufficient evidence of a national-security supply-chain risk and rejected Anthropic's constitutional claims. The ruling allows DOD to continue enforcing the exclusion unless a court overturns it on further review. Anthropic said it was considering its options.
The Big Picture
According to CRS, Anthropic was reportedly willing to modify its usage policies for DOD but refused to authorize mass domestic surveillance or fully autonomous weapons. Anthropic CEO Dario Amodei said, "frontier AI systems are simply not reliable enough to power fully autonomous weapons."
DOD is not publicly known to use Claude or other frontier AI models in autonomous weapons systems. DOD Directive 3000.09, "Autonomy in Weapon Systems," updated in January 2023, distinguishes fully autonomous, human-supervised, and semiautonomous weapon systems. The directive requires appropriate levels of human judgment over the use of force but does not necessarily require manual human control of every engagement.
Congress established reporting requirements through the National Defense Authorization Act for Fiscal Year 2024 (P.L. 118-31) and National Defense Authorization Act for Fiscal Year 2025 (P.L. 118-159). Additional legislation, including the AI Guardrails Act (S. 4113), has been introduced but not enacted.
The Bottom Line
CRS identifies four potential congressional responses without endorsing any: codifying DOD Directive 3000.09, imposing additional reporting or notification requirements, restricting appropriations for autonomous weapons systems, or limiting funding for specific AI-model applications considered too risky. The dispute highlights unresolved questions about congressional oversight, AI reliability, and the military's use of autonomous weapons.
Access the Legis1 platform for comprehensive political news, data, and insights
Spot something wrong? Report an issue with this article