AI leadership sounds impressive until you ask what leadership actually is.
If leadership means scheduling, summarising performance notes, comparing options, preparing talking points, or finding patterns in team data, then yes, AI can help a lot. Honestly, some calendar decisions probably deserve to be automated for the emotional health of everyone involved.
But leadership is not only information processing. It is influence, trust, timing, courage, and the deeply underrated skill of noticing when everyone in the meeting is pretending to agree because lunch is late.
The research summary made this distinction very clear. Employees may be comfortable with AI helping on scheduling, information, budgets, and routine decisions. But they still prefer humans for understanding feelings, building culture, training people, and handling moments where empathy matters. That is not because humans are perfect. Many managers are also powered by caffeine, unread messages, and mild panic. But people still want to believe that the person leading them can understand context.
The AI feedback evidence is especially interesting. In one study, AI feedback improved performance when it was simply deployed, but performance dropped once employees were told the feedback came from AI. That says something important: the content of advice is not the whole story. People react to the source.
There is also the trust problem. Li and Bitterly find that AI management is perceived as less benevolent than human management. In simpler terms: the AI may be competent, but people are not sure it is on their side.
My reflection is this: leaders should absolutely use AI. Use it to prepare better, spot patterns, reduce admin, and become less buried in operational noise. But AI should increase leadership capacity, not replace leadership presence. The leader still has to explain, decide, care, and stand in front of the consequences. Unfortunately, there is no prompt for that part yet.