A senior Goldman Sachs technology leader just put a blunt warning on the record: relying too heavily on artificial intelligence could be quietly wrecking Wall Street’s next generation of talent. Chris Churchman, who leads Goldman’s Marquee platform and co-chairs the bank’s AI working group, says AI reasoning skills are already at risk as algorithms take over more of the analytical work that used to train junior bankers and traders. Here’s what he said, why it matters, and what it means for the future of banking.
Table of Contents
What Did the Goldman Sachs AI Partner Actually Say?
Speaking on Goldman’s internal “Exchanges” podcast, in comments first reported by CNBC, Churchman did not mince words. “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,” he said. It’s a striking admission from inside one of Wall Street’s most AI-forward firms, and it has quickly become one of today’s most talked-about AI reasoning skills warnings.
Churchman compared the risk to how GPS and search engines quietly eroded people’s navigation and memorization abilities over the past two decades. His concern is that the same pattern could hit banking even harder, since so much of a young banker’s judgment is built by grinding through repetitive analytical work under the supervision of more experienced colleagues.
Why AI Reasoning Skills Are at Risk on Wall Street
Junior traders traditionally build instincts by fielding client pricing requests alongside veteran risk takers who can correct mistakes in real time. According to Quartz’s reporting on Churchman’s comments, automating that entire workflow raises an uncomfortable question: “We can absolutely automate that, but then do we get the senior traders that fully understand?” If AI quietly absorbs the repetitive work that once doubled as on-the-job training, the pipeline that turns junior analysts into seasoned decision-makers could break down before anyone notices.
What Is “Cognitive Atrophy” and Why It Matters
“Cognitive atrophy” is the term Churchman used to describe what happens when a skill weakens from disuse, the same way muscles weaken without exercise. Applied to AI reasoning skills, the idea is that constantly outsourcing analysis to a model can leave people less able to reason through a problem from scratch when the model isn’t available, or worse, when it’s confidently wrong.
That risk is amplified in finance because errors are expensive. Churchman noted that Goldman’s own AI platform, Marquee, once volunteered its own limitations unprompted during testing: “It was like, ‘Look, in the end, I’m better at sounding thorough than being thorough.'” That’s a notable admission from the system itself, and it underscores exactly why human reasoning still needs to stay in the loop on high-stakes calls.
How Goldman Is Trying to Protect Junior Bankers’ Skills
Churchman, who ran currency trading at UBS before joining Goldman in 2021, says the firm hasn’t fully “figured out” how to manage the shift yet, but the goal is to design systems where employees still make the final call on consequential decisions rather than becoming passive operators. “You learn by doing, and a lot of knowledge is tacit, it was never written down,” he said, arguing that Goldman needs to preserve that intuitive knowledge in its most experienced people while making sure the next generation still develops it too.
Part of the challenge is technical. Getting AI outputs to be consistently accurate and auditable is, according to Churchman, the hardest engineering problem in bringing AI into Marquee, since finance has far less tolerance for errors than typical consumer AI chatbots. We’ve covered similar accuracy and reliability debates in our breakdown of Google’s Gemini 3.7 Flash model, where the same tension between speed and trustworthiness keeps coming up across the AI industry.
The Bigger AI Reasoning Skills Debate Beyond Wall Street
Goldman is far from alone in wrestling with this tradeoff. As AI infrastructure scales up across every major industry, from the massive server farms behind NVIDIA’s next-generation AI chips to the everyday tools now built into daily consumer tech, the same underlying question keeps surfacing: what happens to human skill when machines handle the thinking?
Researchers outside finance have raised similar flags about over-reliance on generative tools eroding critical thinking in students and knowledge workers alike. Churchman’s comments add a rare, high-profile voice from inside a major bank to that broader conversation, which is part of why the story has spread so quickly since the podcast aired.
What This Means for the Future of Banking Jobs
CNBC has previously reported that Wall Street firms are already exploring ways to reduce the ratio of junior bankers to senior staff using AI, which makes Churchman’s warning about AI reasoning skills feel especially timely. If entry-level roles shrink or change shape, the traditional apprenticeship model that trains future managing directors could look very different within just a few years.
For now, Goldman’s position is that AI should handle repetitive tasks while humans keep ownership of judgment calls, especially in high-uncertainty situations. Whether that balance holds as AI tools keep improving, or whether cost pressure eventually pushes more decisions onto the models themselves, is likely to remain one of Wall Street’s defining AI reasoning skills questions for years to come, as Quartz noted in its own coverage of Churchman’s comments.
Why This AI Reasoning Skills Warning Matters
Coming from a Goldman Sachs partner who helps build the bank’s own AI tools, this warning carries more weight than the usual outside commentary about AI risk. It’s a reminder that even the firms racing hardest to adopt AI are quietly worried about what gets lost along the way. For more on how AI is reshaping tech and business this week, check out our latest 60-second tech roundup.
Frequently Asked Questions
Who warned about AI reasoning skills at Goldman Sachs?
Chris Churchman, a Goldman Sachs partner who leads the bank’s Marquee platform and co-chairs its Global Banking and Markets AI working group, raised the warning during an episode of Goldman’s “Exchanges” podcast.
What is “cognitive atrophy” in the context of AI?
It refers to the idea that reasoning skills can weaken from disuse, similar to a muscle, if people consistently let AI models handle analysis and decision-making instead of reasoning through problems themselves.
Is AI actually replacing junior bankers?
Not entirely, but CNBC has reported that Wall Street firms are exploring ways to reduce the ratio of junior bankers to senior staff as AI takes over more routine analytical work.
What is Goldman’s Marquee AI platform?
Marquee is Goldman’s digital platform for institutional clients, giving hedge funds and other large clients access to the bank’s market data, research, risk analytics, and trade execution tools. Its AI features are currently limited to internal Goldman employees.
How can banks protect employees’ AI reasoning skills?
Churchman suggests keeping humans in charge of high-stakes, high-uncertainty decisions, using AI mainly for repetitive tasks, and deliberately preserving the tacit, hands-on training that has traditionally turned junior staff into experienced decision-makers.
