Goldman Sachs has limited access to advanced AI coding tools, including Anthropic’s Claude Code and Microsoft’s GitHub tools, to engineers under guidance issued in October. The exception is for “strats,” who work across trading, quantitative research, and software engineering.

The bank had allowed bankers and traders to use the tools for months. The change frustrated some traders who had used them to monitor client trading flows, model market moves and positions, and create written and graphic content, a person familiar with the matter said.

A Goldman spokesperson said the bank reviewed productivity and risk management before deciding which employees should use the tools in the long term.

The directive follows an earlier restriction on Anthropic’s Claude for Goldman staff in Hong Kong. That location-specific move came after concerns about exposing confidential information through third-party AI tools.

The change reflects broader caution on Wall Street. Large banks have restricted public coding assistants for production systems. Broker-dealers using AI remain responsible for supervision, recordkeeping, communications, and cybersecurity under existing Financial Industry Regulatory Authority rules.

🔗 Source: Bloomberg

Recent Goldman Sachs developments