Goldman Sachs is expanding its use of artificial intelligence, but one of the executives helping lead that push is warning that too much dependence on AI could weaken the reasoning skills Wall Street needs from its future bankers and traders.
Chris Churchman, a Goldman Sachs partner who leads the bank’s Marquee digital platform for institutional clients, said professionals risk losing the ability to reason from first principles if analytical thinking is increasingly handed over to AI models. He described the potential problem as “cognitive atrophy.”
Why Goldman Sachs is warning about AI
Churchman is also co-chair of Goldman’s Global Banking & Markets AI Working Group, making his comments notable because they come from someone directly involved in the bank’s AI strategy rather than an outside critic.
His concern is not that banks should stop using AI. Instead, he argues that employees must continue understanding problems, building arguments and making judgments themselves rather than simply accepting model-generated conclusions.
The warning arrives as Wall Street puts more capital behind artificial intelligence. Goldman itself is among the financial institutions involved in Nvidia’s planned AI infrastructure financing platforms, illustrating how closely major banks are becoming tied to the technology’s expansion.
Junior bankers may lose valuable hands-on training
One of Churchman’s biggest concerns involves Wall Street’s apprenticeship model. Junior bankers and traders traditionally learn through repetitive analytical work while receiving guidance from experienced colleagues.
AI can complete some of those tasks faster, potentially increasing productivity. But automating too much of the early work could reduce opportunities for younger employees to build intuition and practical knowledge.
Churchman pointed to client pricing requests as an example. Junior traders can develop judgment by handling those requests under the supervision of experienced risk takers. The process can be automated, but removing it raises a difficult question: how will future senior traders gain the same depth of understanding?
AI is already changing banking jobs
The issue is not limited to Goldman Sachs. Banks around the world are restructuring work as automation becomes more capable. Recent Commonwealth Bank workforce changes linked to technology and automation show how AI is increasingly influencing the mix of skills and roles financial institutions need.
For Goldman, that creates a trade-off. AI may reduce routine work and improve efficiency today, while potentially weakening the traditional pathway through which junior employees develop into experienced decision-makers.
Churchman said much of the most valuable knowledge in finance is tacit — learned through experience and interaction rather than written in a manual. Preserving that knowledge means employees still need opportunities to do the work themselves.
HSBC AI Restructuring Puts Banking Jobs and Automation in Focus
AI Sell-Off Tests Investor Confidence Across Global Markets
Accuracy is another challenge for financial AI
Goldman is also dealing with a more immediate technical concern: reliability. In institutional finance, an AI response may need to be not only useful but factual, traceable and auditable.
Churchman said accuracy is one of the toughest challenges in developing AI capabilities for Marquee, which provides institutional clients with access to market data, research, risk analytics and trading services.
Marquee’s AI capabilities are currently available only to Goldman employees. During testing, Churchman said one system effectively acknowledged that it could sound more comprehensive than it actually was — a reminder that confident output is not the same as verified analysis.
Human judgment remains central to Goldman’s AI strategy
Churchman said systems should be designed so employees remain responsible for consequential decisions, especially when uncertainty and risk are high. The goal is to prevent professionals from becoming passive operators who simply approve whatever an AI system recommends.
Goldman has not claimed to have solved that problem yet. In its official discussion on building AI systems for capital markets, the firm explores the need to combine new AI capabilities with the reliability and human judgment demanded by institutional finance.
The significance of Churchman’s warning is therefore broader than a debate about job losses. As AI takes over more routine analytical work, banks will need to decide which tasks should remain part of human training. The productivity gains may be immediate, but preserving reasoning, experience and accountability could determine whether Wall Street continues producing professionals capable of challenging AI when it gets something wrong.
Make Swikblog your go-to source on Google for reliable updates, smart insights, and daily trends.
