The US Defense Department is locked in a high-stakes battle with tech giant Anthropic over the use of powerful artificial intelligence in warfare. The issue goes beyond just “guardrails” and raises questions about how far AI can be used to make life-or-death decisions in military situations.
A senior Pentagon official posed a hypothetical scenario to Anthropic’s CEO Dario Amodei, asking what would happen if an intercontinental ballistic missile was hurtling towards the US with only 90 seconds to spare. The AI system might be able to trigger a response to save the country, but its safeguards wouldn’t allow it. This scenario highlights the tension between the need for military effectiveness and the need to prevent accidents or unethical use of AI.
Source: https://www.bloomberg.com/news/features/2026-02-26/pentagon-pressures-anthropic-to-drop-ai-guardrails-in-military-standoff