'Can't Use AI To Defend From Missile Attack': Why The Pentagon Dropped Anthropic

Pentagon CTO Emil Michael said Anthropic was dropped over restrictions on military AI use.

Advertisement
Read Time: 3 mins
At the centre of the dispute are restrictions surrounding Anthropic's Claude models.
X/@EmilMichael

Pentagon Chief Technology Officer Emil Michael has explained the Defence Department's decision to remove artificial intelligence company Anthropic from its approved technology ecosystem, saying the dispute centred on restrictions the company sought to place on the military use of its AI models.

Speaking on the Ruthless Podcast, Michael said, “Six months into the job, I read the [Anthropic] contract. ‘You can't use AI for battlefield management.' I was like, ‘What does that mean?' ‘You can't use AI to defend yourself from a missile attack.' What?” he said.

Advertisement

ALSO READ: New York Rains: City Faces Severe Thunderstorms, Tornado Risk And Flooding In Low-Lying Areas

At the centre of the dispute are restrictions surrounding Anthropic's Claude models. According to Michael, the company sought to maintain safeguards limiting certain military applications, including autonomous weapons development and battlefield mass surveillance. The Pentagon, however, maintained that technology companies should not determine military policy or operational decisions beyond restrictions established under federal law.

Michael said the Defence Department expects AI providers supplying military networks to permit the use of their systems for all lawful purposes. He argued, “So I called [Anthropic] in, and I was like, 'OK, you guys, you know, got to change this',” Michael told The Fellas. “;We'll use it lawfully, for all lawful purposes, but it's artificial general intelligence. You don't get to tell me where I could fly a plane to. We decide when we buy a plane from Boeing or Lockheed.' And they're like, ‘We think these are our moral values. We have our own constitution.'”

Advertisement

The Pentagon's position became more consequential after other major AI companies agreed to contractual terms allowing their technologies to be used across defence networks. Michael pointed to OpenAI, Google and SpaceX among companies willing to work with the department under its standard requirements.

ALSO READ: Nepal Tragedy: Viral Video Shows Men Stealing Jewellery From Deceased Flash Flood Victim

"Imagine a Chinese hypersonic missile travelling at Mach 5 or faster, giving us just two minutes to intercept it from space. That is far too fast for a human operator with a joystick to handle. Yet, when presented with this situation, Anthropic's response was essentially, 'Call us when it happens, and we'll work it out then'," said Michael.

Essential Business Intelligence, Sharp Market Insights, Practical Personal Finance Advice, Daily Fuel, Gold and Silver Prices and Latest Stories — On NDTV Profit.


Loading...