JPMorganChase blocks Hong Kong staff from using Claude

• Key insight: JPMorganChase has blocked Hong Kong employees from using Anthropic's Claude model.
  
• What's at stake: This incident is part of a growing trend of banks and AI vendors reacting to government concerns about national security.
  
• Forward look: Some experts say other banks should follow JPMorgan's and Goldman's example.

JPMorganChase has cut off its Hong Kong-based employees from Anthropic's Claude AI model, citing licensing and terms-of-service issues, according to the Financial Times. The move follows a similar step Goldman Sachs took in April.

It's part of an emerging trend in which compliance with government orders and concerns about cybersecurity, usage rights, data governance and contractual control are becoming bigger factors in banks' use of frontier AI models.

JPMorganChase and Goldman Sachs declined requests for comment. Anthropic did not respond to a request for comment.

In a blog, Anthropic explained that its terms of service prohibits the use of its products in certain regions "including adversarial nations like China" due to legal, regulatory and security risks. Hong Kong is part of greater China.

In September, the AI model provider said it was strengthening its restrictions in such regions.

"We continue to advocate for policies like strong export controls to prevent authoritarian nations from developing frontier AI capabilities that could threaten national security," the company stated.

Ironically, last week the U.S. government issued an export control that banned the use of Anthropic's Mythos and Fable models by foreign nationals, including Anthropic employees, citing national security concerns. To comply with the order, Anthropic completely blocked use of these models to everyone.

JPMorganChase's removal of Claude from an internal list of approved LLMs that Hong Kong employees could use is "a board-level signal about how banks now negotiate AI tooling," said Oliver Bussman, Bank of America Europe board member and founder and CEO of Bussman Advisory, in a LinkedIn post. "JPMorgan is not blocking Claude globally — the restriction targets a specific region where regulatory scrutiny of cross-border AI deployment is highest."

This distinction "suggests the bank's risk and legal teams determined Anthropic's standard licensing doesn't fit Hong Kong's compliance architecture," he wrote. "In my board work, I see this pattern regularly: institutions moving first on containment rather than waiting for regulators to force the issue."

The question now, according to Bussman, is whether other U.S. and European banks will follow, and whether Anthropic will adjust its terms to unlock institutional access in Asia-Pacific markets.

According to Daniel Wagner, managing director of consultancy Multilateral Accountability Associates, all major U.S. banks should follow the example set by JPMorganChase and Goldman.

"Chase has stopped its staff in Hong Kong from accessing Anthropic's AI models, in a sign of intense scrutiny on the technology's use outside the U.S.," he wrote on LinkedIn.

But according to Theo Lau, founder of consultancy Unconventional Ventures, the recent incidents aren't really about Hong Kong, but much broader.

"Combined with the recent Fable and Mythos shutdown, it is an early indicator of an increasingly fragmented AI world with three blocs: U.S.-aligned, China-aligned, and a hedging middle," Lau told American Banker. "And this has serious and far-reaching implications. A global bank's AI stack is likely no longer going to be global, as some operations will need to shift to locally-deployable models."

AI vendor selection is getting more complex, she said. "It is no longer just a tech decision, but a geopolitical positioning one."