One in three workplace managers has used artificial intelligence, or AI, when writing employee feedback or preparing performance reviews, a survey found.

TechRadar reported on Oct. 1 that a Headway survey of 1,000 managers found 33 percent said they used AI for such tasks.

AI use is spreading beyond employee evaluations to conflict mediation. Some 31 percent said they used AI to resolve team conflicts, while 27 percent said they got AI help preparing for important conversations. Another 57 percent said they wrote scripts with AI or rehearsed in advance ahead of difficult conversations.

Still, the share of AI use in performance reviews fell slightly from a similar survey a year earlier. At the time, 41 percent of managers acknowledged using AI to draft or edit performance reviews. The latest figure is lower, but still significant.

For managers, the incentive to use AI is clear. It can cut time in situations where they must write multiple evaluations each year. The concern is that sensitive feedback that should reflect individual performance and relationships could be left too heavily to automated tools.

Headway burnout prevention and recovery specialist Talia-Maria Tourikis (탈리아-마리아 투리키스) said AI can help organise thoughts. But she said the interaction itself should not be handed over to AI, and pointed out that words written directly by a person come across best in situations that require trust and empathy.

Employees' concerns do not stop at drafting evaluation documents. In a hybrid work survey by Owl Labs of 8,000 workers worldwide, 58 percent were counted as thinking their company had introduced AI-based monitoring tools. That means AI is becoming more prominent in both manager evaluations and workplace surveillance.

The key issue is the level of AI involvement. The latest survey does not confirm whether managers used AI to polish grammar or tweak wording, or whether they used AI to generate the body text of evaluation documents. If performance reviews are tied to salary and compensation, employees could see it as AI, rather than a person, effectively intervening in a process that judges their future.

In such circumstances, the scope of corporate AI use and internal standards are becoming more important. As AI use spreads to evaluations, conflict mediation and preparations for sensitive meetings, organisations may need to clearly define what remains a support tool and what must be left to managers' direct judgement and responsibility.