When the Answer Comes Too Easily: AI and the Slow Erosion of Managerial Judgement

Leaders are hired to think on their feet, to find creative solutions and to weigh the human side of a decision before it is made. That is the quiet premise behind every management layer, and it is the one a new study from the University of Bath puts under pressure. Published in the Academy of Management Review, the research looks at how tools such as ChatGPT may affect the wisdom managers build through real world experience, reflection and human interaction. The question is not whether AI is useful, but what happens to judgement when managers begin to outsource their thinking to it.

What Epistemic Deskilling Means

The researchers use the term epistemic deskilling to describe a steady decline in knowledge related capabilities that follows from overusing AI, as the technology ends up doing much of the thinking. It is a different worry from the familiar one about jobs being replaced. Here the concern is that the people who remain in charge slowly lose the very skills that justified putting them there, without anyone noticing the moment it happens.

Why Convenience Is the Real Risk

The danger lies less in AI’s errors than in its ease. As the article describes it, AI is a tool of convenience rather than detail: it will produce an answer to almost any question, even when it lacks the data or tools to give the right one. Difficult decisions usually have many facets, information to gather and judgements to make, and AI is comfortable presenting results without any of that groundwork. Because the answer arrives so smoothly, future problems are likely to be sent its way too. Over time it becomes easier to accept results without question, and those making decisions may rarely feel fully in control, always turning to the technology for the next step.

What Managers May Stop Doing

Professor Lindebaum, quoted by the University of Bath, describes how this plays out in practice. Managers may stop asking important questions, stop seeking different perspectives and stop learning from real world interactions. Instead of building a nuanced understanding of employees, customers or organisational challenges, they may come to depend on AI generated answers that lack the context and moral judgement complex decisions require. Those are exactly the qualities that make a manager’s call different from an output.

A Wider Pattern

The study sits alongside other warnings. Earlier reports suggest AI could cause workers’ skills to atrophy, and there is evidence that heavy use may flatten individuality, so that everyone’s work starts to echo the same language and ideas. The article also points to a culture of performative AI use, where people try to look proficient in front of colleagues rather than building real skills, often with little direction on how the technology should be used. The common thread is a gradual dilution of the human element of work.

What It Means for Organisations

None of this is an argument against AI, and the article is clear that there is room for balance. It is an argument for being deliberate about which skills and which judgement an organisation wants to preserve as it expands its use of technology. It is a theme close to Euren’s work: assessing how managers and senior leaders actually reason under uncertainty, developing the habit of asking questions and seeking dissent, advising on where AI should inform a decision and where a person must own it, and searching, through executive search, for leaders who can challenge an output rather than accept it. Judgement is built through use. Organisations that want to keep it will need to protect the practice.