Goldman Sachs Restricts Advanced AI Tools to Engineers and Strats
Goldman Sachs has implemented new restrictions on advanced artificial intelligence coding tools, limiting access to specialized technical staff. As reported by Fast Company, the investment bank is prioritizing risk management and productivity by reserving platforms like GitHub and Anthropic’s Claude Code for engineers and quantitative analytics professionals known as strats.
Context from Industry Pulse
The restriction at Goldman Sachs reflects a broader trend among major financial institutions attempting to balance innovation with systemic security. While the bank is limiting advanced coding tools to technical specialists, it is simultaneously piloting autonomous agents and research-focused copilots for its banking divisions. This cautious approach mirrors sentiments shared by executives at JPMorgan Chase and Bank of America, who have highlighted the risks of AI-generated spreadsheets and pitchbooks lacking human verification. As the technology evolves, these firms are shifting away from open-access experimentation toward specialized, role-based deployment to mitigate errors in high-stakes financial environments while still attempting to leverage AI's potential for increased operational efficiency.
The policy shift at Goldman Sachs marks a transition from broad experimentation to a more controlled deployment of high-level software development tools. According to Fast Company, the bank previously allowed non-technical staff to utilize these platforms for creating software prototypes and managing financial data. However, the firm now dictates that once experimental phases conclude, these technologies should remain with the personnel best equipped to manage them for long-term operations.
This decision has reportedly caused friction among the bank’s traders and bankers. Fast Company notes that these employees had integrated AI tools into their daily workflows for tasks such as building financial models, monitoring market data, and generating graphics. The restriction appears to deviate from earlier internal sentiments where leadership praised the ability of non-technical workers to initiate software projects using AI before handing them off to professional developers.
Safety concerns remain a primary driver for institutional caution across the financial sector. Reporting from Fast Company highlights that other major banks are facing similar dilemmas regarding the speed of adoption versus the necessity of human oversight. Executives at peer institutions have recently warned that the rapid evolution of technology often outpaces the development of necessary controls, particularly when AI agents operate without direct supervision.
Despite the new limitations on coding platforms, the firm continues to provide other AI resources to its broader workforce. Fast Company reports that Goldman Sachs maintains a proprietary chatbot for all employees and has been testing specific tools for investment bankers to assist with research compilation. By September 2025, the bank was also piloting autonomous agents designed to support its core development teams, suggesting a tiered strategy for AI access based on job function.
The Industry Pulse Bottom Line
Goldman Sachs is standardizing its AI deployment by centralizing advanced coding tools within technical departments, signaling that Wall Street is prioritizing rigorous risk controls over the widespread democratization of high-level software development tools.