Artificial intelligence company Anthropic and the Trump administration are sparring \u2014 yet again.<\/p>\n
Over the weekend, tensions between the San Francisco company and the Trump administration flared up after the U.S. government ordered Anthropic<\/a> to cut off foreign access to the company\u2019s most powerful AI systems.<\/p>\n The government\u2019s order poses a challenge to Anthropic\u2019s business and its branding, which focuses heavily on AI safety. The company said it disabled access to its most advanced models for all its customers to comply with the directive. Customers still have the ability to access other Anthropic models.<\/p>\n The unusual move came days after Anthropic released the AI models, known as Fable 5 and Mythos 5. Anthropic limited the release of Mythos 5, which can be used to fix software vulnerabilities, to a \u201csmall group of cyberdefenders and infrastructure providers\u201d amid concerns that hackers could exploit the tool to break into computer networks.<\/p>\n Fable 5 excels in software engineering, scientific research and other complex tasks, according to the company. Now, Anthropic is facing security concerns around the release of that model.<\/p>\n Anthropic is competing with other major tech companies to advance AI. Anthropic<\/a> filed to go public and its valuation is nearly $1 trillion, surpassing its rival OpenAI.<\/p>\n Here\u2019s what you need to know about the latest high-stakes public spat:<\/p>\n Why did the U.S. government order Anthropic to cut off foreign access to its latest AI models?<\/p>\n Anthropic said in a blog post on Friday that the government raised national security concerns but didn\u2019t provide details.<\/p>\n \u201cOur understanding is that the government believes it has become aware of a method of bypassing, or \u2018jailbreaking\u2019 Fable 5,\u201d the company said in the post. Anthropic said that there are safeguards in place to prevent the AI system from being misused for cybersecurity tasks and that no tester has been able to broadly bypass these guardrails. <\/p>\n The company added that it worked with the U.S. government ahead of the model\u2019s launch.<\/p>\n But David Sacks, a venture capitalist who served as the Trump administration\u2019s crypto and AI advisor, accused Anthropic of prioritizing \u201cthe continued offering of the consumer model over safety.\u201d<\/p>\n The Trump administration, he said, wants to resolve this safety issue and then access to Fable can be restored.<\/p>\n \u201cThe Admin wants all of this to happen as soon as possible. It is frankly bewildered that Anthropic hasn\u2019t wanted to comply with safety requests that it previously said were its highest priority,\u201d he wrote.<\/p>\n Over the weekend, Trump officials also took jabs at Anthropic, escalating the public feud.<\/p>\n \u201cThree months ago, @DeptofWar kicked @AnthropicAI out of our building \u2014 forever,\u201d Defense Secretary Pete Hegseth said on X. \u201cEvery passing day proves why that was the right move.\u201d<\/p>\n What prompted the government to issue this order?<\/p>\n Conversations that Amazon Chief Executive Andy Jassy had with the Trump administration\u2019s officials including Treasury Secretary Scott Bessent triggered the president\u2019s order, people familiar with the matter told the Wall Street Journal<\/a>.<\/p>\n Amazon researchers were able to get Anthropic\u2019s Fable 5 model to provide information that could be used to help cyberattackers, according to the report. An Amazon spokesperson told the news outlet that \u201cit\u2019s not uncommon for governments to seek our counsel on potential security risks.\u201d<\/p>\n Amazon has also invested billions in Anthropic and supplies them with chips to train their AI models.<\/p>\n Anthropic said it\u2019s complying with the government order but disagrees that the finding \u201cshould be cause for recalling a commercial model deployed to hundreds of millions of people.\u201d<\/p>\n \u201cIf this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers,\u201d the company said.<\/p>\n Have Anthropic and the Trump administration sparred before?<\/p>\n Yes, but it involved a different issue.<\/p>\n