AI overreliance erodes managers' judgment, University of Bath study warns

Generative AI has become as routine as email in many workplaces, but a new study from the University of Bath warns that leaning on it too heavily may come at a hidden cost: managers' judgment. The research cautions that overreliance on Gen-AI can erode the very abilities that make managers effective, including building moral insights, understanding context, and knowing how to get a job done. That is not a warning about a single bad decision. It is a warning about a slow, quiet erosion of the skills that separate a leader from a middleman.

The study, published by researchers at the University of Bath, does not argue that AI is useless. It argues that the way it is used matters. When managers outsource thinking to a chatbot, they may stop practicing the kind of deliberate reflection that builds moral judgment. Contextual understanding, the ability to read a situation, weigh competing interests, and sense what is unsaid, is not something a prompt can generate. And know-how, the practical, hands-on expertise that comes from doing the work, atrophies when a machine does it for you.

This matters because generative AI is no longer a novelty. It is embedded in tools that millions of professionals use daily, from drafting emails to summarizing reports to analyzing data. The convenience is real, and so is the productivity gain. But the Bath study suggests that convenience has a shadow side. If every difficult question gets routed to an AI, the human capacity to answer difficult questions may weaken over time.

The researchers' warning is especially relevant for managers, whose core job is judgment. A manager does not just execute tasks; they decide what matters, who to trust, and when to push back. Those decisions require moral insight, an understanding of right and wrong in a specific situation, and contextual awareness that AI, for all its power, does not truly possess. The study implies that the more managers rely on AI for these judgments, the less capable they may become of making them independently.

For organizations, the stakes go beyond individual performance. If a generation of managers loses the ability to build moral insights, the consequences could show up in compliance failures, ethical lapses, and poor strategic calls. Boards and executives who are pushing AI adoption should consider whether they are also investing in the human skills that AI cannot replace. The study does not say AI is dangerous; it says uncritical use is.

The warning echoes a broader conversation about automation and expertise. In fields from medicine to aviation, research has shown that skills can fade when people rely too heavily on automated systems. The Bath study applies that logic to management, a domain where judgment is the product. The implication is not that managers should avoid AI, but that they should use it as a tool for augmentation, not replacement.

For leaders, the practical takeaway is to build deliberate practices that keep judgment sharp. That might mean setting aside time for reflection, requiring managers to form their own conclusions before consulting AI, or using AI-generated outputs as a starting point rather than a final answer. The study does not prescribe specific steps, but the direction is clear: use AI to inform decisions, not to make them.

The University of Bath study is a reminder that the most valuable asset in any organization is not the technology, but the people who know how to use it wisely. As AI becomes more capable, the premium on human judgment will only grow. Managers who preserve their ability to think, question, and decide will be the ones who lead well in an AI-augmented world.