Two of America’s biggest investment banks, JPMorgan Chase and Goldman Sachs, have reportedly told employees outside the US not to use Anthropic, in a move that highlights rising global scrutiny over advanced AI tools. Both banks have sent a 'Do not use' note on Anthropic to employees outside the US.
According to the Financial Times, JPMorgan has restricted staff in Hong Kong from accessing Anthropic’s Claude models, following a similar step earlier taken by Goldman Sachs. The decision signals growing caution among major financial institutions over how AI tools are governed outside the United States.
ALSO READ: Arizona family left stunned after finding human remains inside urn ordered from Amazon
JPMorgan blocks Claude access in Hong Kong
Employees at JPMorgan in Hong Kong can no longer select Anthropic’s Claude models from the internal list of approved large language models, according to people familiar with the matter.
The move reportedly follows concerns linked to Anthropic’s usage terms and regional restrictions embedded in licensing agreements with corporate clients. One person cited in the report said the decision was influenced by the wording of Anthropic’s contract terms with JPMorgan.
Goldman Sachs had earlier taken a similar approach, restricting bankers in Hong Kong from using Anthropic tools after adopting a strict interpretation of usage rules related to Greater China.
Why Hong Kong is at the center of AI restrictions
While mainland China blocks Western AI tools such as ChatGPT and Claude under the country’s “Great Firewall,” Hong Kong has traditionally maintained broader access to global digital platforms.
However, US-based AI companies often impose their own regional restrictions, especially in sensitive jurisdictions. The latest move reflects how global AI access is becoming increasingly fragmented, with companies applying different compliance rules across regions.
US government tightens oversight on AI exports
The restrictions come amid rising political pressure in Washington over advanced AI models. US Commerce Secretary Howard Lutnick recently sent a letter to Anthropic CEO Dario Amodei directing the company to suspend exports of its advanced models to certain foreign users, citing national security risks.
The directive reportedly includes limits on distributing models such as “Mythos” and “Fable” to foreign nationals without government approval, due to concerns about potential military use in countries like China and Russia.
Europe reacts to growing AI controls
The US move has also sparked concern in Europe, where policymakers fear over-reliance on American AI infrastructure. European leaders are increasingly pushing for “sovereign AI” capabilities to reduce dependency on US technology firms.
French President Emmanuel Macron has led discussions with G7 leaders and tech executives on ensuring access to advanced AI models through trusted partnerships.
German Chancellor Friedrich Merz also emphasized the need for coordination with Washington, saying: “I very much hope that we will have an intensive coordination with the American government, because the potential of these new technologies should be available for all countries.”
The latest restrictions highlight a widening divide in how AI technologies are regulated and distributed worldwide.
As banks, governments and AI companies tighten rules, access to frontier models is becoming increasingly dependent on geography, regulation and security assessments—reshaping how global AI adoption unfolds.
