JPMorgan Chase Bans Hong Kong Staff from Visiting Anthropic, Tightening Compliance Measures Following Goldman Sachs

According to Data Beat monitoring, as reported by sources familiar with the matter to the Financial Times, JPMorgan Chase has halted Hong Kong branch employees' access to the AI model developed by Anthropic. The report notes that Hong Kong branch employees are currently unable to select the Claude model from the internally approved large language model dropdown list. This comes as another Wall Street institution, following a similar move earlier this year by Goldman Sachs, restricts the use of the Claude model in Hong Kong.

Sources indicate that JPMorgan Chase's decision to disable the Claude model is primarily based on the restrictive description of the usage terms in Anthropic's licensing agreement. Previously, Goldman Sachs also underwent a strict interpretation of Anthropic's terms of service. The terms of service explicitly exclude the Greater China region (including Hong Kong) from the scope of service.

The U.S. AI company's restrictions on the Greater China region are partly due to the mitigation of "distillation" risks, where Chinese domestic teams train self-owned large models through frequent calls to cutting-edge models. Currently, Western advanced models such as ChatGPT and Claude cannot be directly accessed in Hong Kong. Multinational corporations typically navigate geographic restrictions by signing global contracts and hosting computation power overseas. However, the restriction measures by Wall Street banks indicate that term barriers arising from geopolitical conflicts are now permeating multinational financial institutions. As the global financial sector widely deploys cutting-edge AI models for core operations such as coding, restricting employee access may have a negative impact on Hong Kong's competitiveness as an international financial hub.