The fate of a $200 million contract hangs in the balance as Defense Secretary Pete Hegseth and Anthropic CEO Dario Amodei clash over AI ethics. But is this a battle for moral high ground or a power play with political undertones? Hegseth is demanding Anthropic relax its safety standards, threatening to cancel the contract unless the AI lab complies by Friday. This ultimatum stems from Anthropic's refusal to allow its AI for domestic mass surveillance and AI-controlled weapons, a stance that has ruffled feathers in the Trump administration.
Amodei has been unwavering in his position, stating that such use of AI is "illegitimate" and open to abuse. But Hegseth argues that the U.S. should be able to utilize AI for all "lawful" purposes, including warfare and surveillance. And this is where it gets controversial—the Trump administration's definition of "lawful" seems to be a bone of contention.
The term "woke AI" has been thrown around by Hegseth and his colleagues, suggesting Anthropic's stance is politically motivated. However, AI experts argue that the term is vague and seems to be a catch-all for any safety measures that restrict the administration's desired use of AI. The White House AI czar, David Sacks, even drafted an executive order targeting tech companies over this very issue.
Other AI companies, like OpenAI and Google, have agreed to more lenient terms, as has Elon Musk's xAI, which was recently approved for classified use. But Anthropic's advanced and secure model made it the Pentagon's first choice for sensitive military applications. Will Anthropic bend to the administration's will, or is this a matter of principle for the AI lab? The clock is ticking as Friday's deadline looms.
