Goldman Sachs Restricts Access to Claude for Hong Kong Staff Amid Rising AI Security Scrutiny

Hong Kong — Goldman Sachs has removed access to Claude, the artificial intelligence system developed by Anthropic, for its bankers in Hong Kong, according to a source familiar with the matter.

The move reflects growing caution among financial institutions over data security, regulatory compliance, and geopolitical risks associated with the use of advanced AI tools.

---

Shift in Internal AI Access

Goldman employees in Hong Kong had previously been able to use Claude through the bank’s internal AI platform. However, access has been withdrawn in recent weeks, while other AI models—including ChatGPT from OpenAI and Gemini from Google—remain available.

Neither Goldman Sachs nor Anthropic has publicly commented on the change.

---

Compliance and Geographic Restrictions

Reports indicate the decision may stem from a stricter interpretation of contractual and compliance requirements, particularly after internal consultations. Notably, Hong Kong is not listed among the officially supported markets for Anthropic’s API and Claude services.

An Anthropic spokesperson has previously stated that Claude has not been formally supported in the region.

---

Broader Context: AI, Data, and Geopolitics

The restriction comes amid heightened tensions between the United States and China over artificial intelligence, data governance, and access to advanced technologies. These issues are expected to feature prominently in upcoming discussions between Donald Trump and Xi Jinping.

Financial institutions are increasingly scrutinizing AI deployments to mitigate risks related to sensitive data exposure, cybersecurity threats, and regulatory compliance—particularly in cross-border environments.

---

Industry-Wide Implications

While it remains unclear whether other banks are implementing similar restrictions, the move highlights a broader trend of tightening controls around AI usage in regulated industries such as banking.

In recent years, leading AI providers have also taken steps to limit access in certain regions. For example, OpenAI has restricted API access in mainland China, citing concerns over misuse and data security.

---

Navigating AI Adoption in Regulated Environments

As financial institutions continue to adopt AI tools to enhance productivity and decision-making, balancing innovation with risk management is becoming increasingly complex.

Goldman Sachs’ decision underscores the evolving challenges organizations face when deploying AI across global operations—particularly in regions where regulatory frameworks, geopolitical considerations, and technology access intersect.