Artificial intelligence has hit Wall Street full force — and there is no denying it is transforming the way banks, traders, and analysts work on a daily basis.
But while everyone is celebrating the efficiency gains, an unexpected warning came from inside one of the most powerful banks in the world.
It was not a tech critic.
It was not a skeptical academic worried about the future.
It was a partner at Goldman Sachs itself — someone leading one of the bank’s top AI projects — who raised the yellow flag. 🚨
The concern is easy to understand but carries deep consequences: what if, by delegating more and more tasks to algorithms, the financial professionals of the future reach the top without ever developing the ability to think for themselves?
That is the paradox at the heart of the discussion Goldman Sachs has brought to the surface — and it is already sparking serious conversations about the role of AI in the financial industry.
Receive the best innovation content in your email.
All the news, tips, trends, and resources you're looking for, delivered to your inbox.
By subscribing to the newsletter, you agree to receive communications from Método Viral. We are committed to always protecting and respecting your privacy.
The Warning That Came From the Inside
Chris Churchman, a Goldman Sachs partner who leads Marquee, the bank’s digital platform for institutional clients, raised a question few expected to hear from someone in his position. Instead of blindly championing automation, he acknowledged a very real risk. In his words, there is a huge danger that in the age of AI, people will outsource their own thinking to these models and end up suffering a kind of cognitive atrophy, losing the ability to reason from first principles.
The comment was made during an episode of the Exchanges podcast, produced by the bank itself, and the transcript was shared exclusively by CNBC. The problem Churchman pointed out is not the technology itself — it is the passive dependence on it, without building a critical foundation behind it. He made a spot-on comparison: just as people have lost navigation and memorization skills because of modern inventions, bankers risk losing their analytical abilities if they let algorithms do all the heavy lifting.
Think about it: when you use a GPS your entire life and never learn how to find your way on your own, what happens when the device dies in the middle of a trip? In the financial world, that analogy is frighteningly real. A junior analyst who never had to build a financial model from scratch, because AI always delivered it ready-made, could have serious trouble questioning a flawed assumption, spotting an inconsistency in the data, or simply understanding why a result does not make sense. And on Wall Street, mistakes like that are expensive — sometimes extremely expensive.
Churchman made it clear that reasoning is still essential. According to him, you still need to think through problems and structure them into a coherent argument, and the current risk is precisely that we are delegating that reasoning to machines. What makes this warning even more significant is exactly who delivered it. Churchman is not an outside observer with a skeptical view of technology. He is inside Goldman Sachs, leading AI projects, betting on this transformation. So when someone with that profile says it is necessary to reflect on the impacts on human development, the message carries a different weight. It is not resistance to change — it is responsibility toward the future. 💡
The Dilemma Playing Out on Wall Street
Wall Street’s push to integrate artificial intelligence into every trading and investment banking process could be a kind of risky bargain. On one side, it makes the industry more profitable today. On the other, it could erode the very talent it will need tomorrow. With AI taking over more and more of the routine work that traditionally taught young bankers and traders how to think and make decisions, institutions risk sacrificing the culture that turns entry-level employees into seasoned professionals.
There is even a more direct consequence in this story: the potential reduction in the need for junior bankers. CNBC had previously reported that Wall Street firms were exploring ways to use AI to reduce the ratio of junior professionals relative to senior ones. In other words, beyond the question of developing reasoning skills, there is also a real impact on career structure within these banks.
Financial analysis has always been one of the most demanding activities in the market. It involves interpreting complex data, understanding the macroeconomic context, assessing risks that often are not spelled out in any spreadsheet, and making decisions that affect billions of dollars. Historically, shaping a strong analyst took years — and much of that learning happened precisely through the repetitive and seemingly simple tasks at the beginning of a career.
If those entry-level tasks are completely absorbed by artificial intelligence, how will professionals build the repertoire they need to make decisions in ambiguous situations? Financial markets are not predictable. Crises emerge without warning, contradictory data pops up, and the best decisions often require an intuition that only comes from accumulated experience. Experience that needs to be built — not outsourced to an algorithm. 📉
Learning Culture at the Center of the Debate
Churchman, who ran the foreign exchange desk at UBS before joining Goldman Sachs in 2021, argues that banks need to find a balance between using AI and preserving the learn-by-doing culture that has long defined Wall Street. According to him, much of the knowledge in this industry is tacit and intuitive — things that were never written down anywhere. You learn by doing.
In his view, Goldman Sachs needs to make sure that the intuitive knowledge the bank’s best professionals carry today does not get lost. And more than that, it needs to ensure the next generation develops it too. A practical example he gave was about junior traders, who learn by handling client pricing requests under the supervision of experienced risk operators. It is entirely possible to automate that task, Churchman acknowledges. But then the uncomfortable question arises: will we still be developing senior traders who truly understand what is going on?
The most mature answer to this dilemma is not about simply limiting the use of AI — that would be naive and impractical. It is about building a learning culture that coexists with automation. That means ensuring that even with AI handling operational tasks, professionals continue to be exposed to the reasoning behind each process, understanding the fundamentals and developing the ability to question the results algorithms deliver.
Churchman also highlighted a crucial design point: systems need to be built in a way that keeps employees in control of high-risk, high-uncertainty decisions, rather than turning them into passive operators. And he was honest in admitting that not even Goldman Sachs, one of the largest investment banks in the world, has figured out how it will manage this transition that has already begun. Churchman, by the way, also co-chairs the AI working group within the bank’s Global Banking and Markets division. 🧠
The Quest for Error-Free Answers
On the same podcast, Churchman shared valuable lessons about implementing AI in Marquee, the platform used by hedge funds and other institutional clients to access market data, research, risk analytics, and trade execution services from Goldman Sachs. For now, the AI-powered version of Marquee is available only to the bank’s employees.
From a technical standpoint, the biggest challenge is ensuring that AI responses are one hundred percent factual and auditable. While consumer-facing chatbots warn users about potential errors, in the world of high finance the tolerance for mistakes is extremely low. A single piece of incorrect information can trigger decisions that move enormous sums of money.
Churchman shared a curious episode from the platform’s development. When challenged firmly, the AI itself made a surprising admission. It acknowledged, with a certain honesty, that it is better at seeming thorough than at actually being thorough. That confession from the machine perfectly captures the core of the concern: language models can sound extremely convincing even when they are not being accurate. And that is where the human ability to question and verify becomes irreplaceable.
A New Model for Growing With AI
What Goldman Sachs is signaling is that the financial industry needs to rethink how it trains its professionals in a world where artificial intelligence is already part of the work environment from day one. This is not a problem exclusive to Wall Street — it is a question that will come up in virtually every field where AI is being adopted at scale. But in the financial market, where the consequences of flawed reasoning can be catastrophic, the urgency is greater.
The discussion Churchman sparked opens an interesting path: instead of treating AI as a replacement for human reasoning, viewing it as a tool that amplifies the reasoning skills of those who have already developed them. In this model, the technology enhances human talent — it does not replace or weaken it. For that to work, though, organizations need to invest in people just as much as they invest in tools.
At the end of the day, the warning that came from inside Goldman Sachs is less about the technology and more about choices. Choices about how to integrate AI into daily professional life, how to structure the training of the next generation of analysts, and how to make sure the efficiency gained today does not create a silent fragility tomorrow. These are questions that financial institutions — and many other industries — will need to answer carefully in the coming years. And the sooner that conversation truly begins, the better for everyone. ✅
