The U.S. Department of Commerce has directed AI company Anthropic to restrict access to its latest “Claude 3” models for all non-U.S. citizens, according to a statement released Tuesday. The order, which affects foreign employees and international collaborators, including researchers in India and the European Union, is framed as a measure to prevent “unauthorized dissemination of sensitive AI capabilities,” according to a Commerce Department spokesperson. Anthropic confirmed the suspension in a public filing, stating it is “cooperating fully with federal authorities.”
The restriction applies to access for non-U.S. nationals at both Anthropic and partner organizations, according to a person familiar with the matter who spoke on condition of anonymity. The move aligns with broader U.S. efforts to regulate advanced AI systems amid concerns over their potential military or surveillance applications. A Commerce Department official declined to specify which models are impacted beyond “the latest generation of Claude,” citing ongoing investigations.
India’s Ministry of Electronics and Information Technology confirmed it is monitoring the situation, with a spokesperson stating, “We are in active dialogue with U.S. counterparts to understand the implications for cross-border AI collaboration.” Indian tech firms, including Tata Consultancy Services and Wipro, have previously partnered with Anthropic on enterprise AI projects, according to public records. Similar restrictions have been reported against other U.S. AI firms, including OpenAI and Meta, though those companies have not publicly acknowledged such measures.
The policy shift underscores intensifying geopolitical tensions over AI development. In 2023, the U.S. imposed export controls on advanced semiconductor equipment to China, citing national security risks. Analysts note that AI systems now face parallel scrutiny, with the Commerce Department’s Bureau of Industry and Security overseeing rules that could restrict data flows and computational resources. A draft memo from the bureau, obtained by The Wall Street Journal, outlines potential criteria for classifying AI models as “critical technologies,” though no official rules have been finalized.
Anthropic’s decision to comply with the order has drawn mixed reactions. A group of European AI researchers issued a statement calling the move “premature and counterproductive,” arguing that “international collaboration is essential to address shared risks like algorithmic bias and safety standards.” The European Commission has not commented directly on the restriction but has previously expressed concerns about U.S. tech regulations hindering transatlantic innovation partnerships.
The U.S. government has not provided a timeline for when access might be restored. Anthropic’s public filing notes that the company is “exploring legal and procedural avenues” to challenge the order, though no formal appeal has been filed. The firm’s CEO, Dario Amodei, previously warned in a 2023 interview that “regulatory overreach could stifle the very innovation that makes AI beneficial to society.”
Industry observers are tracking how the policy evolves, particularly its impact on global AI research. The National Institute of Standards and Technology, which advises the Commerce Department, is conducting a review of AI risk frameworks, according to a federal calendar. A final report is expected by the end of 2024, though the agency has not confirmed whether the Anthropic case will influence its findings.
