Goldman Sachs partners warn: Outsourcing reasoning to AI may lead to the next generation of bankers experiencing a "cognitive crisis."
A partner at Goldman Sachs, who is leading the firm's core artificial intelligence project, has issued a warning: as AI rapidly infiltrates Wall Street, there is a risk that the independent thinking ability of the next generation of financial professionals may be weakened.
A partner at Goldman Sachs who is leading the bank's core artificial intelligence project has issued a warning: As AI rapidly permeates Wall Street, the independent thinking abilities of the next generation of financial professionals are at risk of being diminished. Chris Churchman, head of Goldman Sachs' institutional client digital platform Marquee, stated, "In the age of AI, we are likely to outsource all reasoning processes to these models, leading to a decline in cognitive abilitieswe will lose the ability to reason from first principles."
Churchman pointed out that just as modern inventions have gradually caused people to lose their navigation and memory skills, if algorithms take over all heavy analysis, bankers' analytical and judgment capabilities may also degrade.
"Reasoning is still crucial," he emphasized. "You still need to unfold each question and construct it into a logical argument, and we are now handing this process over to AI."
Wall Street is fully embedding AI into trading and banking processes, which could create a "devilish trade-off": it might enhance industry profits in the short term but could erode the talent foundation needed for the future. As AI takes over a significant amount of the routine work that traditionally trained junior bankers and traders, jobs that were their "training grounds" for thinking and decision-making, the loss of these roles may dismantle the apprenticeship culture that has shaped newcomers into seasoned Wall Street talent.
The long-term implications may be even more severe, as AI could directly reduce the need for junior bankers. Reports from last year indicated that Wall Street firms are exploring how to use AI to lower the ratio of junior to senior employees.
Churchman believes banks must find a balance between the use of AI and the traditional apprenticeship model. Before joining Goldman Sachs in 2021, he was in charge of foreign exchange trading at UBS.
"Knowledge is acquired through practice, and a lot of skills are tacit knowledge that has never been formally documented," he said.
Goldman Sachs must "ensure that we do not lose that kind of tacit intuitive knowledge that our top talents currently possess, and that the next generation can also master," Churchman noted.
He gave an example that junior traders typically gain experience by responding to client quote requests under the guidance of senior risk decision-makers.
"We can fully automate this process," he said, "but then can we still cultivate truly seasoned traders who understand the entirety of the market?"
System design must ensure that employees retain the final decision-making authority in high-risk, high-uncertainty decisions, rather than becoming passive operators, Churchman stressed.
Even as a top global investment bank, Goldman Sachs has not yet "found the answer"how to manage this ongoing transformation, Churchman admitted. He also serves as co-chair of the bank's AI working group for the Global Banking & Markets division.
The "zero-error" dilemma
In the interview, Churchman also shared insights from integrating AI into the Marquee platform. Marquee provides hedge funds and other institutional clients with Goldman Sachs' market data, research, risk analysis, and trading execution services.
He indicated that currently, the Marquee AI platform is only available to Goldman Sachs employees.
From a technical perspective, the biggest challenge lies in ensuring that the answers generated by AI are 100% accurate and auditable. Consumer-grade AI chatbots like Siasun Robot & Automation often warn users that they may make mistakes, but in high-end finance, the margin for error is extremely slim.
Churchman revealed that during the development of the customer-facing AI platform, the software candidly stated a warning: "When we pressed it hard, at least it was honest," Churchman said. "It responded: 'At the end of the day, I am good at sounding comprehensive, rather than being truly comprehensive.'"
Related Articles

Bessembinder's former mentor sharply criticized: Buying back U.S. Treasury bonds is a mistake! Defiance of fundamentals keeps yields low and will "ultimately be defeated."

The second draft of the Agricultural Law amendment aims to improve measures that support the development of agriculture and the rural economy.

Musk: The launch date for the first batch of AI satellites equipped with Nvidia chips has been set.
Bessembinder's former mentor sharply criticized: Buying back U.S. Treasury bonds is a mistake! Defiance of fundamentals keeps yields low and will "ultimately be defeated."

The second draft of the Agricultural Law amendment aims to improve measures that support the development of agriculture and the rural economy.

Musk: The launch date for the first batch of AI satellites equipped with Nvidia chips has been set.

RECOMMEND





