Goldman Sachs executive warns AI could create “cognitive atrophy” across Wall Street

Goldman Sachs is urging caution about the financial industry’s accelerating use of artificial intelligence, with a senior executive arguing that the technology could dull people’s thinking rather than simply improve efficiency. Chris Churchman, head of the Marquee digital platform at Goldman Sachs, warned that AI-driven workflows could lead to what he described as “cognitive atrophy” on Wall Street.

In remarks highlighted in a recent Yahoo Finance report, Churchman framed the issue as a “huge danger” and suggested the problem is not only technical, but human. His core concern is that increased reliance on AI systems may weaken the mental discipline required for tasks like evaluating risk, interpreting information, and making independent judgments.

Churchman’s comments come as firms across finance explore AI for a wide range of uses, from research and customer support to market analysis and internal automation. For Goldman, the warning also has a practical edge: it comes from the person responsible for Marquee, which the report describes as Goldman’s digital platform serving institutional and corporate clients.

While the report centers on Churchman’s warning, it does not, in the information provided here, spell out specific incidents, internal metrics, or policy changes at Goldman related to the “cognitive atrophy” risk. It also does not identify which AI systems or workflows he believes are most vulnerable to the human-skills erosion he described.

Marquee, as characterized in the report, sits at the intersection of Goldman’s technology efforts and client-facing services. Digital platforms like this typically bundle software capabilities, connectivity, and workflow tools aimed at reducing friction for large customers. In that context, Churchman’s comments imply that product design and operational controls matter, not just the underlying AI model.

The broader sector backdrop is that AI adoption in capital markets is moving quickly, and competitive pressure can push firms to integrate capabilities faster than governance frameworks can mature. In that environment, a warning about “overreliance” reads as a call for stronger oversight of how AI recommendations are used in day-to-day decisions.

A key limitation of the available reporting is the lack of disclosed detail on what Goldman Sachs is doing specifically to address the risk. The report does not provide specifics such as new guardrails, model monitoring standards, training changes, or quantitative targets. As a result, investors and clients will likely watch for follow-on statements that connect Churchman’s concerns to concrete internal controls and client usage policies.

Going forward, the issue to monitor is whether Goldman translates this warning into operational changes that limit blind reliance on AI outputs. That could include clearer human-in-the-loop expectations, tighter risk validation around AI-generated analysis, and more transparent documentation of how clients should interpret and challenge recommendations.