Anthropic comes under Pentagon's radar over AI safety

Anthropic CEO Dario Amodei confirmed that the Pentagon has blacklisted the AI startup after refusing to comply with its demands. The decision, which labels Anthropic’s Claude AI as a supply chain risk, bars any company with defence contracts from working with the startup.

Despite this, Amodei claimed that they have had “productive conversations” with the Department of War, which is ready to take legal action against the designation.

Amodei said that Anthropic have been looking for ways to support the Pentagon while still abiding by their ethical boundaries. However, Undersecretary of Defence Emil Michael denied any active negotiations with the AI company, stating on X that “there is no active @DeptofWar negotiation with @AnthropicAI.”

The official designation blacklists the AI company, which means they cannot be contracted by the government. However, Amodei claimed that “the vast majority” of their enterprise customer base is unaffected by this designation.

Microsoft, which uses Anthropic AI models on their platforms, such as M365, GitHub, and AI Foundry, has confirmed that they will still be supplying their customers with Claude AI, excluding defence-related projects.

Moreover, Amodei also issued an apology to the public for his earlier memo that was highly critical of the White House, stating that it no longer represents his views. Amodei went out of his way to state that Anthropic is committed to de-escalation, writing that “Anthropic has much more in common with the Department of War than we have differences.”