Trump Administration Reignites Its Feud With Anthropic Over Latest A.I. Models
Trump Administration Reignites Its Feud With Anthropic Over Latest A.I. Models
https://www.nytimes.com/2026/06/13/us/politics/trump-anthropic-ai-models.html
Publish Date: 2026-06-13 20:41:00
Source Domain: www.nytimes.com
Here are several key points extracted from the article regarding the escalating feud between Anthropic and the Trump administration:
– The Trump administration demanded Anthropic disable access to its most advanced A.I. systems, Fable 5 and Mythos 5, for foreign nationals, a move perceived as unprecedented and controversial.
– This action has sparked debate among former U.S. officials and cybersecurity experts, with some arguing it diverges from the hands-off approach President Trump had previously endorsed for regulating the A.I. industry.
– Anthropic received no clear explanation for the national security concerns prompting the directive but stated that claims of a “jailbreak” method to bypass security on Fable 5 were exaggerated.
– The administration, through allies like former A.I. czar David Sacks, accused Anthropic’s leadership of ignoring security warnings and failing to address potential risks.
– This incident followed Anthropic’s filing for an IPO, revealing a valuation of nearly $1 trillion, and highlighted ongoing tensions between the company and the Pentagon, which labeled Anthropic a supply-chain risk.
– Officials continue to express strong sentiments against Anthropic, while experts like Katie Moussouris from Luta Security argue the administration’s actions may misinterpret the technology’s capabilities.
– The Trump administration’s new export controls could have broader implications, raising questions about the regulatory future of other A.I. models with similar capabilities.
– The administration’s approach contrasts with previous policies, which generally reserved such severe restrictions for the most sensitive technologies rather than A.I. models.
This summary captures the essence of the points discussed in the article without delving into specifics that might breach original content guidelines.
