US Banks Restrict Anthropic AI Access in Hong Kong
Two of America's largest investment banks, JPMorgan Chase and Goldman Sachs, have instructed employees outside the United States to avoid using Anthropic's AI models. According to a report by the Financial Times, JPMorgan Chase has blocked its Hong Kong staff from accessing Anthropic's artificial intelligence systems, following a similar move by rival Goldman Sachs. This reflects heightened scrutiny over the use of such technology beyond US borders.
Employees at JPMorgan in Hong Kong can no longer select Claude models from the internal drop-down list of approved large language models (LLMs). Sources familiar with the matter told the Financial Times that the decision was based on the wording of Anthropic's usage terms in its licensing agreement with JPMorgan. Earlier this year, Goldman Sachs also prohibited its bankers in Hong Kong from using Anthropic, interpreting the company's terms strictly to exclude usage in Greater China.
Context of Restrictions
Western AI models like OpenAI's ChatGPT and Anthropic's Claude are banned in mainland China due to the Great Firewall, which restricts external information. However, Hong Kong has historically operated with minimal Chinese censorship. The current restrictions in Hong Kong are imposed by US AI companies themselves. While direct access to models is blocked from Hong Kong, international organizations have circumvented geographic limits by negotiating global contracts and hosting activity outside China.
US Government Orders Export Suspension
Last week, US Commerce Secretary Howard Lutnick sent a letter to Anthropic CEO Dario Amodei, ordering the company to halt exports of its Mythos and Fable AI models worldwide and to all foreign nationals. The order cited concerns that these models could be used by military intelligence users in China, Russia, and other countries of concern. Without a license from the Commerce Department, Anthropic must not provide its Fable 5 and Mythos 5 models to foreign nationals anywhere, due to national security risks.
European Alarm Over US AI Ban
The US order to limit foreign access to cutting-edge AI models has alarmed Europe, which worries about its dependence on American technology. A movement is growing in Europe to develop sovereign AI capabilities. French President Emmanuel Macron led talks with Group of Seven leaders and tech executives on deploying advanced AI through trusted partners. German Chancellor Friedrich Merz stressed the need for coordination with the US government to ensure global access to these technologies.
During G7 discussions, Anthropic CEO Dario Amodei and OpenAI CEO Sam Altman called for international collaboration on AI safety initiatives. Altman emphasized balancing safety with global access and benefits. Anthropic has previously advocated for a system where governments and developers jointly decide when to slow AI progress to mitigate risks.
