Nielsen: AI Makes Employee Evaluations More Difficult

    Although the artificial intelligence tools companies provide to employees speed up work, they create new challenges for managers. Nielsen conducted a survey in 7 countries. A total of 1,700 directors and senior managers participated in the study. More than half of the executives said that evaluating employees’ work has become more difficult since AI was introduced. This is because managers looking at the finished product cannot distinguish which parts of the work were done by the employee and which were done by AI.    Despite this, 94% of executives have a positive view of using AI to improve workflows. About half of managers spend more time than before reviewing and editing materials prepared by their teams. The problem is not that AI tools perform poorly. Companies’ internal rules are not changing as quickly as the technology. As a result, tools that speed up employees’ work can increase managers’ workloads.    Among survey participants, 78% have changed the way they evaluate employees since AI was introduced. Writing effective prompts for AI, deciding which tasks to assign to it, and experimenting with new ways of working are now considered more important skills. Of the managers, 46% use AI to help them make decisions. Another 39% of executives use the technology to prepare written evaluations of employees’ work.    Using AI in employee evaluations increases the risk of bias. Large language models—systems that generate text based on existing data—can reproduce the human biases present in their training data. A study published in early 2025 found that AI models can be biased against Black male candidates. Companies have yet to establish clear rules addressing these risks.    The adoption of AI has also not reduced the number of software tools used by companies. Of the managers, 57% said their organizations now use more tools than before. According to 51% of respondents, software tools have become more interdependent. Only 29% consider their organizations’ technical systems for AI to be fully organized and centrally managed. Employees’ use of AI tools that have not been approved by the company makes oversight even more difficult.    Of the managers, 58% expect to oversee more AI agents than human employees over the next 1 year. AI agents are systems that carry out assigned tasks with a certain degree of autonomy. Of the respondents, 46% are concerned about losing control over their actions. Another 35% are worried about being held responsible for decisions made by AI. Executives therefore want these agents to operate under human supervision rather than with full autonomy when handling important tasks.