Photo: Wikimedia Commons (archive photo)
Goldman Sachs has limited access to some advanced coding tools to engineers after a period of broader employee experimentation, according to internal guidance updated in October and reviewed by Bloomberg News. The change affects tools used by bankers and traders, but is not a bank-wide prohibition on artificial intelligence.
The guidance restricts Anthropic’s Claude Code and Microsoft’s GitHub to engineers, Bloomberg reported in an article published by Moneycontrol on Monday. Goldman had allowed bankers to use the tools for work for several months before narrowing access.
Some traders were frustrated by the change, a person familiar with the matter told Bloomberg. They had used the tools to monitor client trading flows, model changes in markets and positions, and produce written and graphic material.
Productivity and risk review
Goldman spokesperson Nick Carcaterra told Bloomberg that the bank makes selected AI tools available to employees to identify useful applications across its businesses. After experimentation, he said, the firm assesses factors including productivity and risk management before deciding which employees are best suited to use those tools over the long term.
That explanation describes a narrower allocation of particular products, rather than an abandonment of AI. Goldman has developed its own assistant combining models from OpenAI, Anthropic and Alphabet, Bloomberg reported.
The shift also follows public enthusiasm for nontechnical staff building with AI. As recently as July, Chief Information Officer Marco Argenti had described traders and bankers developing solutions with AI tools and then passing those creations to engineers for improvement, according to Bloomberg.
A boundary around coding tools
The October guidance draws a boundary between experimenting with code-generating products and deciding who should have ongoing access inside a financial institution. Bloomberg’s report did not describe a ban on every external AI service or every use of AI by non-engineers.
The bank’s review reflects the practical questions raised when software-building tools move beyond engineering teams: whether the resulting work improves productivity and how its risks are managed. Carcaterra’s statement identified those considerations but did not disclose the details of Goldman’s internal assessment.
For bankers and traders who had incorporated the tools into everyday tasks, the reported change means access now depends on their role. For Goldman, the broader AI effort continues, with the restrictions applying to the specified coding tools rather than the firm’s full set of AI applications.
JBizNews Desk | New York
© JBizNews.com. All rights reserved. Unauthorized reproduction or redistribution is strictly prohibited.
