Goldman Sachs (GS) reportedly restricts access to advanced AI coding tools, including Anthropic's Claude Code and Microsoft's (MSFT) GitHub, to engineers only, Bloomberg reports

Goldman Sachs (GS) reportedly restricts access to advanced AI coding tools, including Anthropic's Claude Code and Microsoft's (MSFT) GitHub, to engineers only, Bloomberg reports.

Bank-internal restrictions on third-party AI tools fit an established pattern: financial institutions have historically been slow adopters of external software that touches code or client data, with compliance, data-leakage and model-governance concerns typically driving gating to controlled user groups before any broader rollout. The read-across here is less about the named vendors than about the sequencing banks have tended to follow, pilot with engineers, build audit and data-controls around the tool, then widen access, so a restriction headline is often a staging step rather than a rejection. For the vendors, enterprise deals with large banks have usually been sticky once embedded, and access-scoping decisions at a single customer have not historically been a reliable signal on contract status. Worth noting is the distinction between a security-driven control decision, which is routine at regulated firms, and a strategic pullback on AI spend, which would be newsworthy; nothing in the headline separates the two. Follow-ons are any vendor comment, whether peer banks announce similar scoping, and whether the restriction extends beyond coding tools to other AI products. As a single-source report on internal policy, the signal is low-conviction.