Source: Li & Bitterly (2024) — “How Perceived Lack of Benevolence Harms Trust of Artificial Intelligence Management,” Journal of Applied Psychology
An AI manager can be efficient.
It can schedule. Track. Allocate. Evaluate. Remind. Measure. Optimise. Basically everything a workplace loves to turn into a dashboard.
But this paper points to a problem that dashboards rarely solve: benevolence.
Across field and experimental evidence, the authors find that AI management is perceived as less benevolent than human management. That matters because trust in a manager is not only about ability. It is also about whether the person, or system, is believed to care about your interests.
This is where AI management gets awkward.
A human manager may be imperfect, biased, tired, distracted, or replying to messages at a worrying hour. But employees can still imagine that the manager understands their situation. Maybe they care. Maybe they can make an exception. Maybe they can see context.
AI management feels different.
Even if it is accurate, it can feel indifferent. It may apply rules consistently, but consistency is not the same as care. If the system denies a request, changes a schedule, or evaluates performance, the employee may experience it as cold rather than fair.
That is a serious adoption problem.
Especially in management, because people accept difficult decisions more easily when they believe the decision-maker has at least considered their welfare. AI may be consistent, but consistency without care can feel like bureaucracy with better software.
Because people do not only ask, “Is this manager capable?”
They ask, “Is this manager on my side at all?”
If the answer feels like no, trust suffers.
My takeaway: AI management cannot rely on competence alone. Organisations need to design for perceived care, appeal, context, and human backup. Otherwise AI may become the manager that never forgets a metric but never understands a person. Efficient, yes. Trusted, not necessarily.