A Better Newspaper

## Overview Goldman Sachs has stopped bankers in Hong Kong from using Anthropic's Claude AI models, with employees reportedly unable to access the company's AI tools as of a few weeks prior to reporting (FT). The move signals emerging caution among major financial institutions about deploying third-party generative AI tools in sensitive jurisdictions. ## Key Details - Employees at Goldman Sachs in Hong Kong were unable to access Claude models as of a few weeks before the report (FT). - The restriction reportedly applies specifically to the Hong Kong office, raising questions about regional regulatory, data security, or geopolitical considerations. - Anthropic's Claude models have been widely adopted in professional services, making enterprise restrictions like this a significant signal for the AI vendor landscape. ## Why It Matters This is an early example of a major financial institution restricting access to a leading AI lab's models in a specific jurisdiction, potentially reflecting concerns about data sovereignty, cross-border information flows, or geopolitical risk tied to US-China tensions. For attorneys and compliance professionals, this signals a growing wave of enterprise AI governance decisions shaped by jurisdiction-specific legal and geopolitical risk, particularly in Hong Kong given its unique regulatory position between China and international finance. ## Developing Angles - Whether other banks or financial institutions in Hong Kong or Asia follow suit. - Anthropic's response and any public statements about the restriction. - Broader implications for AI vendor selection amid US-China regulatory friction. - Comparison to other reported AI restrictions (e.g., Bank of England review, NSA policy reversals) tracked elsewhere.