AI Is Reshaping Finance — But Human Intelligence Still Runs the Show

The financial world is awash in artificial intelligence optimism. From JPMorgan to Visa, major institutions are repositioning themselves as AI-first technology companies rather than traditional money handlers. Yet beneath the bold proclamations, a more sobering picture is emerging — one that reveals a growing disconnect between the promise of AI and its practical impact on the ground.

A new survey from the Judge Business School at Cambridge University, drawing on responses from 628 finance firms, AI companies, and regulators across the globe, has laid bare the uncomfortable truth: AI is largely confined to back-office functions, and more than three-quarters of large financial institutions struggle to even measure its value. Only 40 percent of firms reported any profit improvement from AI adoption, while 43 percent reported no change.

Compare that to an Nvidia-commissioned survey claiming that 89 percent of finance executives say AI is boosting revenues, and the credibility gap becomes impossible to ignore.

The 'AI Native' Problem Nobody Is Talking About

One of the most telling signs of AI's uneven integration comes not from boardrooms, but from internship programs. A prominent New York financier recently shared a candid observation: his 2025 summer intern cohort was the first he had encountered that he would call "true AI natives" — individuals who had grown up immersed in both digital technology and artificial intelligence from the start.

The result? Impressive on the surface, hollow underneath. When senior staff probed the depth of their ideas, the work fell apart. The firm subsequently reduced return offers and began steering recruitment toward humanities graduates — students trained in critical reasoning, contextual thinking, and nuanced judgment.

The lesson is clear: AI can generate outputs. It cannot yet generate wisdom.

The Regulatory Blind Spot

While the private sector sprints ahead, regulators are being left behind. The Cambridge survey found that regulators' AI adoption rates are roughly half those of the institutions they oversee — a dangerous asymmetry in a sector where speed and systemic risk go hand in hand.

The risks are real and multiplying. The Financial Stability Board has flagged concerns over the surge of private credit financing of AI data centers, warning of likely defaults. The International Monetary Fund has raised the specter of systemic cyber attacks. Other documented threats include algorithmic "herding" — where AI-driven systems pile into identical trades simultaneously — over-reliance on a handful of cloud providers operating outside regulatory reach, and model hallucinations producing flawed financial decisions.

Responses are beginning to emerge. In the United States, Treasury Secretary Scott Bessent and Federal Reserve Chair Jay Powell have convened meetings with top financiers to assess emerging threats. In the United Kingdom, the Financial Conduct Authority is offering fintech firms free computing power and data access within a supervised "supercharged sandbox"—a controlled environment designed to enable responsible innovation without systemic exposure.

Automation Is Not the Enemy — Misuse Is

The pattern playing out in finance is one that automation advocates have long anticipated. Technology does not deploy itself wisely. It amplifies whatever strategic intelligence — or lack thereof — guides it.

"AI is not a replacement for human judgment — it is a multiplier of it. The organizations that will win are not those with the most AI tools, but those with the most capable people who know how to direct them. That is exactly why I built the Automation Institute: to close the gap between access to technology and the ability to use it meaningfully." — Hamza Baig, Founder, Automation Institute & Hexona Systems

This perspective cuts to the heart of what the Cambridge data is telling us. Across surveys from McKinsey, EY, the Institute of International Finance, and others, the gap between AI rhetoric and AI results is consistent and persistent. One survey from technology firm Hyland found that only 45 percent of businesses say AI is delivering the outcomes they expected.

The remaining 55 percent are not failing because AI is inadequate. They are struggling because they deployed AI without the human infrastructure to support it.

What the Finance Sector Must Do Next

The data points toward a clear strategic imperative. Financial institutions — and indeed organizations across every sector — need to invest not just in AI tools, but in developing people who can interrogate, direct, and quality-control those tools.

Notably, the Cambridge survey found that only a quarter of financial firms expect AI to result in sector-wide job losses, while 58 percent actually anticipate that AI will drive more hiring or reskilling within their own organizations. This is not the apocalyptic narrative of mass displacement. It is a reskilling challenge — and an enormous opportunity for those positioned to meet it.

The firms, educators, and regulators who understand this distinction will define the next era of finance. Those who mistake access to AI for mastery of it will find themselves exposed.

As attention now turns to the class of 2026 — the next wave of interns and graduates entering financial services — the central question is no longer how much AI they have used. It is whether they have developed the human judgment to use it well.