Goldman Sachs exec does not ‘agree’ with banks using AI models of OpenAI, Anthropic and other, says: We are delegating what is very important to us

The rapid adoption of artificial intelligence (AI) across Wall Street risks undermining the analytical capabilities of future financial leaders, according to a senior Goldman Sachs executive directing one of the investment bank’s premier digital and AI initiatives. Speaking on a recent episode of the firm’s Exchanges podcast, Chris Churchman, a Goldman Sachs partner who heads the institutional client platform Marquee and co-chairs the bank’s Global Banking and Markets AI working group, cautioned that outsourcing fundamental problem-solving to automated models threatens core human reasoning.“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. Reasoning is still important. You still need to reason about [problems] and structure it into an argument, and now we’re delegating reasoning,” Churchman stated in remarks shared with CNBC.

The threat to apprenticeship and Wall Street culture

According to Churchman, just as everyday navigation and memorisation eroded with the modern navigational tools, financiers risk losing their core critical instincts if software performs all complex evaluations. This dynamic creates a double-edged sword for the financial sector: while deploying AI increases immediate profitability, it risks hollowing out the foundational skills junior staff need to develop into senior decision-makers.Churchman, who led foreign exchange trading at UBS prior to joining Goldman in 2021, highlighted that institutional expertise relies heavily on hands-on practice and unwritten, intuitive insights passed down across trading desks.While routines like handling client pricing requests can be entirely automated, bypassing those entry-level tasks raises concerns over whether future traders will fully grasp risk management. He also said that the reliance on AI also threatens entry-level headcount, following reports last year that Wall Street firms were exploring ways to use automation to reduce the proportion of junior analysts relative to senior staff.Next, for Churchman, systems must ensure staff members make the final judgment calls during high-stakes, ambiguous scenarios instead of acting as passive software monitors.Churchman also discussed the practical engineering realities of embedding AI into Marquee, Goldman’s proprietary platform used by hedge funds and institutional clients for trade execution, research, market data, and risk analytics.