Goldman Partner Warns AI Could Erode Wall Street Talent Pipeline

The apprenticeship problem

A senior Goldman Sachs executive overseeing one of the bank's core artificial intelligence initiatives has raised concerns that widespread AI adoption could undermine how Wall Street develops its future leaders.

Chris Churchman, a partner who leads Goldman's Marquee digital platform for institutional clients, warned that delegating analytical work to algorithms may produce "cognitive atrophy" among junior employees who would traditionally build reasoning skills through hands-on tasks.

"There's a huge danger here that in the era of AI, we outsource our reasoning to these models, and we have cognitive atrophy that stops us being able to reason from first principles ourselves," Churchman said during Goldman's "Exchanges" podcast.

Why it matters

Wall Street's apprenticeship culture has long transformed entry-level employees into experienced decision-makers through repetitive, supervised work. If AI handles those formative tasks, banks face a paradox: short-term efficiency gains that may hollow out the talent pipeline needed for long-term leadership. The challenge extends beyond Goldman—any knowledge industry relying on learning-by-doing must grapple with how automation reshapes skill development.

The tacit knowledge challenge

Churchman, who previously ran currency trading at UBS before joining Goldman in 2021, emphasized that much of Wall Street expertise exists as unwritten, intuitive knowledge passed down through direct experience.

Junior traders typically learn by handling client pricing requests under the watch of senior risk managers. "We can absolutely automate that," Churchman acknowledged, "but then do we get the senior traders that fully understand?"

The concern mirrors how modern tools have already eroded other cognitive capabilities. Just as GPS reduced navigation skills and smartphones diminished memorization, Churchman suggested that AI could weaken the analytical muscles financiers need for complex decision-making.

Churchman, who also co-chairs Goldman's Global Banking and Markets AI working group, admitted the firm hasn't yet "figured out" how to manage this transition despite already being deep into implementation.

Technical hurdles in high-stakes finance

Beyond workforce development, Churchman outlined technical challenges in deploying AI for institutional clients through Marquee, Goldman's platform providing market data, research, risk analytics and trade execution.

The platform's AI capabilities remain limited to Goldman employees for now. The primary obstacle is ensuring 100% factual accuracy with full auditability—a far higher bar than consumer chatbots that simply warn users about potential errors.

During development, Churchman said the AI system made a revealing admission when pressed on its limitations: "It was like, 'Look, in the end, I'm better at sounding thorough than being thorough.'"

That gap between appearing authoritative and being accurate represents a fundamental challenge for AI deployment in high finance, where tolerance for mistakes remains extremely low.

Churchman stressed that systems must be designed to keep humans in control during high-stakes, high-uncertainty decisions rather than reducing employees to passive operators of automated tools.

The details were first reported by CNBC, which obtained an exclusive transcript of the podcast interview.