Goldman Sachs Restricts AI Coding Tools for Traders and Bankers, Limits Access to Engineers

Goldman Sachs Group Inc. has restricted the use of advanced AI coding tools among most of its employees, limiting access primarily to engineers after reviewing productivity and risk management considerations.

For several months, the bank had allowed bankers and traders to experiment with Anthropic’s Claude Code and Microsoft’s GitHub tools to automate tasks and develop solutions for their day-to-day work. However, according to internal guidance updated in October and seen by Bloomberg News, access is now being restricted to engineering teams.

Traders Express Frustration Over Restrictions

The decision has reportedly frustrated some traders who had been using AI coding tools to improve their workflows.

According to a person familiar with the matter cited by Bloomberg, employees had used these tools for activities such as monitoring client trading flows, modelling market changes and positions, and creating written and graphical content.

The restrictions mark a shift from the bank’s earlier approach of encouraging non-technical employees to experiment with AI-assisted software development.

Goldman Had Encouraged ‘Vibe Coding’

As recently as July, Goldman Sachs Chief Information Officer Marco Argenti had highlighted the potential of “vibe coding” among non-technical employees, including traders and bankers.

Under this approach, employees use AI tools to develop applications and workflow solutions without requiring extensive programming expertise. Their creations can then be handed over to the bank’s engineering teams for further development and refinement.

Argenti had said that employees were developing solutions using AI tools and passing them on to one of the firm’s approximately 12,000 engineers to improve.

The latest restrictions suggest that the bank is reassessing how such experimentation should be managed across different business functions.

Goldman Cites Productivity and Risk Management

Goldman Sachs spokesperson Nick Carcaterra said the bank makes selected AI tools available to employees to identify suitable applications across different parts of the organisation.

He explained that after an experimentation period, the bank evaluates several factors, including productivity and risk management, before deciding which employees are best positioned to use the tools over the long term.

The approach reflects the challenges financial institutions face in balancing the potential efficiency gains from AI with operational controls and technology risks.

Wall Street Banks Reassess Enterprise AI Usage

Goldman’s decision comes as major financial institutions increasingly invest in AI while developing policies around its deployment.

Banks are spending millions of dollars on AI technologies but are also evaluating concerns related to cybersecurity, implementation costs, productivity and the appropriate use of different AI models.

JPMorgan Chase, for instance, said in July that it was encouraging employees to determine which AI model was best suited to specific tasks. The objective was to avoid using more expensive models for relatively simple activities, such as summarising an analyst report.

Goldman Sachs has also developed an internal AI assistant that combines models from OpenAI, Anthropic and Alphabet.

AI’s Growing Role in Goldman Sachs’ Business

The bank’s position on AI is particularly significant as it continues to engage with companies driving the technology boom.

Goldman Sachs has been named a lead underwriter for Anthropic’s planned initial public offering and has partnered with chipmaker Nvidia to arrange major debt transactions, according to the supplied report.

In May, John Waldron, widely viewed as a potential successor to Chief Executive Officer David Solomon, described Goldman Sachs as a “human assembly line” with significant potential for automation.

The bank has also previously restricted employees’ use of Anthropic’s Claude in Hong Kong, indicating that its AI policies may vary across locations and business functions.

Restrictions Do Not Cover Quantitative Technology Specialists

The latest directive does not apply to so-called “strats” employees who work at the intersection of trading, quantitative research and software engineering according to another person familiar with the matter cited by Bloomberg.

Goldman’s latest move highlights a broader shift in enterprise AI adoption: financial institutions are moving beyond experimentation towards more controlled deployment, with access increasingly determined by job responsibilities, productivity assessments and risk management requirements.

By : Vanshika Tayal

Picture of Indian BFSI Times

Indian BFSI Times

Leave a Reply

Your email address will not be published. Required fields are marked *