Key Takeaways
- Anthropic’s Claude AI has been deactivated for Goldman Sachs banking personnel in Hong Kong within the past few weeks
- Alternative AI platforms such as ChatGPT and Gemini continue to function on Goldman’s internal systems
- Following a comprehensive contract examination, Goldman determined Hong Kong staff should discontinue all Anthropic product usage
- Anthropic’s official platform does not designate Hong Kong as an authorized service region
- Financial regulators and banking institutions globally are examining Anthropic’s latest Mythos model for potential banking sector vulnerabilities
Investment banking powerhouse Goldman Sachs has discontinued access to Anthropic’s Claude artificial intelligence platform for its Hong Kong-based workforce, Reuters learned from a directly informed source.
The insider revealed to Reuters that banking professionals in Hong Kong previously enjoyed Claude access via Goldman’s proprietary AI infrastructure. This capability was revoked in recent weeks.
Competing AI solutions, including Google’s Gemini platform and OpenAI’s ChatGPT service, remain accessible to personnel through the identical internal system.
The Financial Times initially disclosed this development, referencing sources with knowledge of the situation. Goldman Sachs refused to provide commentary. Anthropic has yet to respond to media inquiries.
The FT reported that Goldman arrived at this conclusion after conducting a rigorous interpretation of its contractual agreement with Anthropic. The financial institution engaged in discussions with Anthropic prior to determining that Hong Kong-based staff should cease utilizing any Anthropic offerings.
An Anthropic representative informed the FT that Claude had never received official “support” designation for Hong Kong operations. The company’s digital platform excludes Hong Kong from the roster of territories where its API infrastructure or Claude.ai services are formally accessible.
The Hong Kong Factor
Hong Kong occupies a unique position regarding American AI technology access. While mainland China prohibits AI platforms developed by U.S. corporations, Hong Kong has maintained relative openness, with accessibility determined by individual technology providers.
This development unfolds against escalating U.S.-China friction concerning artificial intelligence capabilities and information security. An anticipated summit between Presidents Donald Trump and Xi Jinping scheduled for mid-May is projected to address these issues.
The United States government released a worldwide advisory last week regarding purported AI intellectual property theft by Chinese AI entities. During 2024, OpenAI implemented measures to restrict Chinese network traffic to its API infrastructure, citing concerns about potential model exploitation by Chinese competitors for training purposes.
Reuters was unable to verify whether additional financial institutions or corporations have imposed similar Claude restrictions in Hong Kong.
Mythos Model Under Regulatory Microscope
Goldman’s decision coincides with heightened scrutiny of Anthropic’s most recent AI system, Mythos, from international banking institutions and financial oversight bodies.
The Hong Kong Monetary Authority confirmed to Reuters that it has reached out to numerous prominent banks to assess recent developments surrounding Mythos. The regulatory body additionally instructed these institutions to refresh their risk evaluations and implement suitable protective measures.
Reuters documented that certain lending institutions were already evaluating additional security protocols in response to cybersecurity challenges associated with sophisticated AI models.
During February, Goldman Sachs Chief Information Officer Marco Argenti disclosed that the bank was collaborating with Anthropic to create AI-driven autonomous agents designed to automate an expanding spectrum of operational tasks.
The Hong Kong government failed to provide commentary on Reuters’ inquiry. The HKMA refused to discuss Goldman’s particular determination.


