Goldman Sachs cuts AI access: Why Hong Kong Bankbars lost Claude

Goldman Sachs has removed access to AI models developed by Anthropic for its bankers in Hong Kong, according to a source familiar with the matter.

Staff in the city had previously been able to use Anthropic’s Claude models through the bank’s internal AI platform. However, access has been withdrawn in recent weeks following a review of contractual terms, News.Az reports, citing Reuters.

The move reportedly stems from Goldman Sachs adopting a stricter interpretation of its agreement with Anthropic after consultations with the company, leading to a decision that employees in Hong Kong should not use its AI products.

Despite the restriction, other major AI tools — including models from OpenAI and Google — remain available to staff via the same internal system.

While AI platforms developed by U.S. companies are generally restricted in mainland China, Hong Kong has largely operated under different rules, with access policies often set by the companies themselves.

An Anthropic spokesperson said its Claude models had never been officially supported in Hong Kong, without providing further details. Goldman Sachs declined to comment on the changes.

The decision does not affect the bank’s broader partnerships with other AI providers, even as it continues exploring the use of AI-powered tools to automate internal operations.

The development highlights growing complexity around AI governance, cross-border access, and corporate compliance as financial institutions expand their use of advanced technologies.

News.Az 

By Aysel Mammadzada