The United States lifts its restrictions on Anthropic’s most powerful AI models
The United States Department of CommerceThis Tuesday it lifted restrictions on all Anthropic artificial intelligence models,allowing the company to get its technology technologies back on trackMore powerful. . .
The United States Department of CommerceThis Tuesday it lifted restrictions on all Anthropic artificial intelligence models,allowing the company to get its technology technologies back on trackMore powerful AIaccording to a government letter accessed by The New York Times, addressed to the company. In the letter, the Secretary of Commerce, Howard Lutnick, informs Anthropic thatyou no longer need a licensefor exports or transfers within the country of its artificial intelligence modelsClaude Mythos y Claude Fable,thus revoking an order issued on June 12. “Anthropic has taken measures in close coordination with the United States governmentto face the risksassociated with Claude Mythos 5 and Claude Fable 5,” Lutnick wrote. The move marks the latest chapter in Anthropic’s often rocky relationship with the federal government and is intended tomitigate the conflictbetween both parties, paving the way for the company to largely recover its usual activity. Citing national security concerns, the Commerce Department had initially orderedthat Anthropic suspended the access of any person of foreign citizenshipto its most recent artificial intelligence models. That limitation raised questions about the degree of intervention the Trump administration could exert over the rapidly growing artificial intelligence industry. It could also have caused considerable problems for the technological development of Anthropic, which, along with OpenAI, is amongthe most influential AI companies in the world. Mythos and Fable, very effective fordetect security flaws in software,They are considered a significant improvement over previous AI models. The restrictions mark the second time the Trump administration has targeted Anthropic, whose headquarters are inSan Francisco, California. In March, after tense talks about how Anthropic’s artificial intelligence could be used in war,The Pentagon called the company an unacceptable risk to its supply chainwhich could limit its use by federal agencies. Following this assessment,Anthropic sued the federal government. Lutnick lifted export controls after weeks of deliberations between the Commerce Department and the company over safeguards built into artificial intelligence models. Last week the Secretary of Commerce announced a partial agreement with Anthropic toreset accessto its Mythos model for some clients. In Tuesday’s letter, Lutnick notes that Anthropic agreed to take certain steps that include proactively detecting and addressing security risks to its AI models,work in collaboration with the US government on protocols and standards for these and future models,and report any malicious activity to the government. Lutnick added that the Commerce Department “reserves the right to reconsider the decisions made in this letter” if Anthropic does not meet its commitments or if circumstances change. Through a statement published on social networks, Anthropic announced that it would beginrestore access to its AI models this Wednesday. “We thank our users for their patience and all those who collaborated with us in the redistribution of the models,” the company stated there. Some time ago, President Trump had adopted anon-interference stanceregarding the artificial intelligence industry, arguing that itIt had to grow unhindered for the United States to compete with China. However, in what represents his most significant step towards regulating this technology, this month he signed an executive order that asked technology companiesvoluntarily allow the government to oversee new AI modelsbefore releasing them to the public. Since then, the Trump administrationproceeded to regulate other artificial intelligence companiesof Silicon Valley, in addition to Anthropic. Introducing an AI technology called GPT-5. 6 Sol this monthOpenAIannounced that it was only going to share it with those companies approved by the administration. Last week,Meta— the only major American AI development corporation that has not reached an agreement to voluntarily allow the federal government to review its models — was pressured by the administration to start sending it its models. (The New York Times has sued OpenAI and Microsoft forcopyright infringementon informative content related to AI systems. Both companies denied such allegations. )According to two people familiar with the discussions, Trump officials are still working on a framework for how companies should formally submit new AI models for review and what standards they should adhere to. In a message posted on social media on Tuesday, White House Chief of Staff Susie Wiles expressed her “gratitude to the companies from various industry sectors that continue to collaborate closely with the White House. ”to implement Trump’s executive order on artificial intelligence. “Our shared priority remainsimplement the best technology as quickly and safely as possible”, he communicated.