The impact of AI: business schools are the canary in the coalmine

There is a reassuring way to talk about artificial intelligence in business education. It says that AI is an important new topic to be incorporated into the portfolio; schools must adjust their curricula; executives need to understand its implications; students must learn to use it responsibly; and faculty must experiment with new tools.

Leading schools are making AI more central, recognising that it cannot be treated as a peripheral technical subject or left to engineers and computer scientists alone. It is becoming part of the context in which strategy, finance, marketing, operations, entrepreneurship, governance and organisation must be understood.

But AI is much more than a topic. It is also a shock to the business school model itself. Business schools are not outside the economy looking in. They are an unusually exposed version of the pressures that will increasingly be felt across the knowledge economy. In that sense, they may be the canary in the coalmine.

Consulting firms, law firms, advertising agencies, investment banks, corporate strategy departments, policy units and professional service partnerships all rely on people who can gather information, frame a problem, compare options, synthesise evidence, generate a point of view and present it persuasively. Gen AI can do this quickly and cheaply enough to alter the economics, perceived scarcity and status of the activity. It changes not only how work is done, but where value is located.

Business schools should recognise this dynamic because we are implicated in it. We trade in structured understanding. We teach concepts, frameworks, cases, models and modes of analysis. We help students and executives become more strategically literate, more organisationally aware, more financially conversant and more capable of engaging with complexity.

At our best, we deepen judgement and broaden managerial imagination. But we should also admit that not everything we have historically rewarded has been deep judgement. Much of it has been the ability to produce a plausible analysis of a curated situation, expressed in the appropriate vocabulary and delivered in a recognisable format. AI is already quite good at that.

Devaluation, not disappearance

This does not mean that AI will replace business schools, just as it will not simply replace lawyers, consultants, managers or analysts. The issue is not disappearance but devaluation and reconfiguration. When a capability becomes easier to imitate, its price falls, its prestige weakens, and the basis of differentiation moves elsewhere.

If AI can produce a reasonable case summary, a competent market map, a strategy memo, or a comparison of business models, the question for business schools is whether enough of what we do can remain scarce, defensible and consequential. The relevant comparison is not between AI and the finest examples of human judgement. It is between AI and the average output that organisations actually tolerate, circulate and reward.

AI systems hallucinate, overgeneralise and miss context – but so do people. AI can produce generic recommendations, but much managerial discourse is generic. AI can be overconfident, but the same could be said of many strategy decks, student essays and executive presentations. The danger is not that machines have become wise. The danger is that a significant amount of what passes for managerial understanding was never as demanding as we liked to think.

The business school story therefore matters beyond business schools. AI exposes the vulnerability of activities built on codified expertise, symbolic fluency and repeatable formats of advice. It shows what happens when the language of competence becomes easier to reproduce. A consulting firm that sells slides without implementation has a problem. A law firm that sells standardised drafting without distinctive judgement has a problem. A marketing agency that sells fluent content without deep customer insight has a problem.

Analysis becomes less distinctive when it’s easy to produce. Other abilities become more important. Framing problems, grasping context, testing assumptions. Being able to mobilise others. Being willing to take responsibility for action under uncertainty. These are harder to automate – not because they are mystical human attributes, but because they are embedded in situations, relationships, incentives, power structures and consequences.

Fluent output and noisy signals

AI not only makes some forms of output easier to produce. It also makes many existing signals noisier. A polished essay, a confident presentation, a competent memo, a neat analytical framework or a careful-looking market scan may tell us less than before about the underlying capability of the person or team that produced it. This matters in teaching, but it also matters in hiring, promotion, professional services, managerial evaluation and institutional credibility.

Business schools have always depended on signals. Degrees signal capability. Grades signal mastery. Case performance signals judgement. Executive education certificates signal development. These signals are imperfect, but serviceable enough to support institutional routines. AI makes these signals less reliable. If essays, case analyses and managerial recommendations become easier to generate, a competent output no longer suffices. What matters is whether it reflects understanding, originality, standards, responsibility and the capacity to act.

This creates a sharper challenge than the familiar concern about cheating. Of course academic integrity matters. But the deeper issue is whether our assessments and credentials still measure what we claim to value. Has the task itself become an inadequate proxy for learning? Likewise, if managers can produce more fluent strategy documents, the question is not whether the documents sound convincing. It is whether they help the organisation make better choices.

AI therefore exposes the difference between fluency and understanding. Fluency is the ability to produce language that fits the expected form. Understanding is the ability to know what matters, why it matters, what follows, what might be wrong, and what should be done. As fluent output becomes cheaper, we will need to become much more precise about what we teach, what we assess and what we certify.

The benefits will not be evenly distributed

The benefits of powerful technologies are rarely distributed evenly. They tend to flow disproportionately to individuals and organisations already well positioned to use them, and to those willing to adjust their value proposition around the new technology. Those who understand that AI requires a wholesale repositioning will benefit as others struggle. The technology will augment some, commoditise others and redefine competitive dynamics.

This matters for business schools. In a world of abundant, fluent and often unreliable output, we need places where claims about AI, productivity, work, skills, organisations, regulation and value creation can be tested with independence and discipline. We need research that distinguishes adoption from impact, experimentation from transformation, productivity theatre from genuine productivity, and automation from value creation. The more noise there is, the more valuable impartial judgement becomes.

This may also intensify hierarchy among schools. Institutions that combine research credibility, employer relationships, alumni networks and live access to practice may become stronger, while those built around generic content and credentialing may become more exposed.

What comes next

What happens when plausible analysis, fluent synthesis and polished output become abundant, and real value now resides elsewhere? Business schools face this question sooner and more visibly because they trade in knowledge, analysis and credibility. But the same question will confront consultancies, law firms, corporate centres, banks, policy units, creative agencies and many other professional organisations.

When fluency is cheap, institutions must become clearer about what they uniquely develop, certify and contribute. That is the uncomfortable but useful warning. AI has not made serious education, research or advice less important. It has made their foundations harder to take for granted.

What’s next for business schools

• AI should be treated as a managerial, organisational and strategic phenomenon, not simply as a technical topic.

• The case method and experiential learning need to move beyond elegant diagnosis towards judgement under uncertainty.

• Schools must reconsider what they certify when competent-looking analysis can be generated on demand.

• Their broader role is to curate serious knowledge in a noisy AI environment, distinguishing hype from evidence and adoption from impact.