Sachs Group Inc. has irked some of its staff after it banned most of them from using advanced artificial intelligence tools designed for coding after reviewing productivity and risk concerns.

For months, Goldman allowed bankers to use Anthropic PBC’s Claude Code and Microsoft Corp.’s GitHub for work. Now, the bank is limiting the use of those tools to engineers only, according to internal guidance updated in October and seen by Bloomberg News.

The ban has frustrated some traders, who had used the tools to accelerate aspects of their work such as monitoring client trading flows, modeling market changes and positions, and creating written and graphic content, according to a person familiar with the matter, who asked not to be identified talking about private information.

As recently as July, Goldman Chief Information Officer Marco Argenti touted the “vibe coding” by non-technical employees at Goldman, such as traders and bankers. Those staff members had developed solutions with AI tools and passed on their creations to one of the firm’s 12,000 engineers to improve, he said in an interview.

“We make certain AI tools available to our people to determine the best use cases across different parts of the firm,” said Nick Carcaterra, a Goldman spokesperson. “After a period of experimentation where we assess a variety of factors including productivity and risk management, we put them in the hands of those who are best suited to leverage them for the long term.”It’s the latest twist as Wall Street irons out just exactly how it wants its thousands of workers using AI models. Banks are spending millions of dollars on AI tools while also trying to figure out how to navigate the increasing concerns around cybersecurity risk, costs and productivity.

JPMorgan Chase & Co. said in July that it was seeking to have employees figure out what model to use for what purpose, helping to avoid using the most expensive model for a task as simple as summarizing an analyst report. Goldman itself has developed an AI assistant that combines models from OpenAI, Anthropic and Alphabet Inc.

At the same time, Wall Street firms are benefiting from the surging financing needs of AI firms. Goldman has been named lead underwriter on Anthropic’s blockbuster initial public offering and has partnered with chipmaker Nvidia Corp. to arrange major debt deals.

In May, John Waldron, the heir apparent to Chief Executive Officer David Solomon, said he viewed Goldman as a “human assembly line” that was ripe for automation.

It’s not the first time that the bank has pulled back on allowing employees to use certain AI tools. Earlier this year, the bank restricted the use of Anthropic’s Claude by its staff in Hong Kong, a location-specific restriction.

The bank’s latest directive doesn’t apply to so-called “strats,” employees who sit at the intersection of trading, quantitative research and software engineering, said another person.