The decision removes Claude from the bank’s internal roster of approved large language models, according to sources familiar with the matter. While Hong Kong has historically maintained fewer internet restrictions than the mainland, Western AI firms are increasingly tightening their geographic boundaries. Anthropic maintains that its models were never officially supported in the territory, placing the burden of compliance on corporate users operating under global enterprise agreements.
In section Cryptocurrency
JPMorgan Blocks Employee Access to Anthropic’s Claude in Hong Kong
JPMorgan Chase has cut off internal access to Anthropic’s Claude models for employees based in Hong Kong. The move follows a similar restriction at Goldman Sachs, driven by Anthropic’s licensing terms that explicitly exclude the Greater China region from its service area, forcing global banks to re-evaluate their local AI infrastructure.
This shift highlights the growing friction between the rapid integration of frontier AI in financial workflows and the tightening regulatory landscape. Beyond regional restrictions, Anthropic faces mounting pressure on multiple fronts. The company recently disabled its Fable 5 and Mythos 5 models following a U.S. government directive regarding export controls and potential software vulnerability risks. Simultaneously, the firm is navigating a proposed class-action lawsuit in California, where subscribers allege that premium Claude plans fail to deliver the usage capacity promised in marketing materials. These developments coincide with Anthropic’s own public push for more rigorous, government-mandated cybersecurity standards for the most advanced AI systems.
Comments (0)
No comments yet. Be the first!