Chatbots and management

If artificial intelligence is more empathetic than a manager

The potential of chatbots highlights the importance of awareness of managerial biases to improve workplace relationships

9 May 2025

Aggiungi Il Sole 24 Ore
ai preferiti su Google

4' min read

More and more often, in work contexts, people report relational difficulties with their manager. Without going as far as extreme cases of mobbing, more or less voluntary, there are everyday situations that can generate deep damage: burnout, requests for psychological support, talent drain. Stifling micro-management, veiled prejudices, passive-aggressive jokes, lack of genuine listening are just some of the most visible dynamics. But alongside these, other more subtle ones lurk: the fear of asking questions, the anxiety of not being up to scratch, feeling isolated or ignored compared to other colleagues. Phenomena that, if neglected, compromise not only individual well-being but also the overall quality of organisations.

In this scenario, the adoption of chatbots based on generative artificial intelligence could represent a concrete answer to a relational void that many workers experience on a daily basis. The question then becomes inevitable: might not a chatbot in certain situations prove to be more helpful, more competent, more humble and more empathetic (albeit through a simulation) than many managers in the flesh? A provocation, perhaps. But a provocation that is borne out by theoretical observations and scientific evidence.

Managerial bias: ubiquitous and under-recognised

Numerous studies confirm that cognitive biases pervasively influence managerial judgement and behaviour. The halo effect (attributing positive value to an employee just because he/she is brilliant in a single area), confirmation bias (looking for signs that confirm an initial impression of unreliability), implicit favouritism (favouring those with a background or communication style similar to one's own), primacy effect (allowing oneself to be disproportionately influenced by first impressions, positive or negative), proximity bias (paying more attention and recognition to those who are physically closer, e.g. those who work more often in attendance) are well documented in the behavioural and decision-making psychology literature.

According to Harvard Business Review, 82% of employees believe that they do not receive objective evaluations from their direct managers. This finding is accompanied by another significant observation: the relational quality of management is one of the main determinants of turnover. In other words, people do not leave the company, they leave their boss.

In reality, while the presence of bias in human beings is universally recognised, managers often lack a real awareness of such biases, and thus an inclination to address them systematically.

Better fake empathy than genuine disinterest

On the flip side, it is true that Gen AI can also inherit bias from the data it is trained with. However, unlike the human being, this vulnerability is explicitly recognised in the design of the models and is subject to continuous corrective action. Artificial intelligence is born, so to speak, already aware of its own fallibility.

Finally, a curious fact emerges, which for me deserves to be analysed: a chatbot does not possess real empathy, but it can simulate it, and in doing so, it often succeeds in ensuring more coherent, neutral and respectful communication than many human managers who simply do not exercise it.

It will have happened to you too, at least once, to exchange a few jokes with a virtual assistant - perhaps with ChatGPT - and to feel the pleasant sensation of being listened to, welcomed, even encouraged. Here, know that you are not alone. A study published in Computers in Human Behaviour showed how people feel more listened to and less judged when confronted with empathetic conversational chatbots, compared to human interlocutors. The main reasons? Absence of judgement, predictability of language, assertive tone, active listening. It is not surprising then that, as noted in a recent article published in Harvard Business Review, the main spontaneous use that users make of artificial intelligence is for psychological support, both in the personal and business spheres.

When the chatbot is preferable to the manager

It is not a question of replacing human relations with algorithms, but of recognising that in many everyday situations the chatbot might be preferable. For example:

- in providing neutral feedback

- in supporting the analysis of priorities

- in handling difficult conversations

- in listening without fear of being judged

Under these circumstances, the chatbot's communicative consistency and neutrality become a relational competitive advantage, especially in environments where managers are under pressure, poorly trained or lack emotional intelligence. An increasing number of workers, without explicitly saying so, would rather relate to a virtual assistant than to their direct manager. This is not so much because of an overconfidence in technology, but because of an accumulated disappointment in human relations.

Conclusions

Artificial intelligence is not the solution. But it is a mirror. An accelerator of awareness. If it shows us that many workers feel more comfortable with a virtual interface than with a flesh-and-blood manager, it is not the technology we need to question. It is us. The way we build power, trust and listening in organisations. And perhaps - paradoxically - it is precisely an artificial assistant that will help us to finally become more human.

*Corporate Manager .