00:00Anthropic will not bend to the Pentagon's demands.
00:03Defense Secretary Pete Hegseth gave the company until Friday evening
00:07to provide the military with unrestricted access to its AI model, Claude.
00:11If it refused, the Pentagon warned it could cancel a $200 million contract
00:16and label the company a supply chain risk,
00:19or even invoke the Defense Production Act to force compliance.
00:23Still, CEO Dario Amadei held firm.
00:27These threats do not change our position.
00:30We cannot in good conscience accede to their requests,
00:33given the substantial value that Anthropic's technology provides to our armed forces,
00:38we hope they reconsider.
00:40Amadei says the Defense Department failed to address his main concerns,
00:44that Claude could be used for mass surveillance or fully autonomous weapons.
00:49Sean Parnell, the Pentagon's top spokesperson,
00:52dismissed these concerns in a social media post Thursday.
00:56The Department of War has no interest in using AI to conduct mass surveillance of Americans,
01:03nor do we want to use AI to develop autonomous weapons that operate without human involvement.
01:09This narrative is fake and being peddled by leftists in the media.
01:14Currently, Claude is the only AI model available for the military's classified systems.
01:20In light of the standoff, the Pentagon is looking to speed up negotiations with OpenAI,
01:25Google, and XAI.
01:28Those companies have already agreed to loosen safeguards for use on unclassified systems.
01:33However, they are not yet used for more sensitive classified work,
01:37and it remains unclear if their AI models are capable of replacing Claude.
01:42Claude's الله coming to the next question.
01:42Thank you, sir.
Comments