Overreliance on Gen AI tools at work may impair critical thinking skills: Microsoft report

The report named “The Impact of Generative AI on Critical Thinking: Self-Reported Reductions in Cognitive Effort and Confidence Effects From a Survey of Knowledge Workers” examines two key questions: when and how knowledge workers perceive the inaction of critical thinking when using Gen AI; and when and why do they perceive increased or decreased effort for critical thinking due to Gen AI. The researchers studied 319 knowledge workers and examined 936 real-world examples of AI use. 

What does the research say?

The research was conducted by Hao-Ping (Hank) Lee from Carnegie Mellon University, in collaboration with Advait Sarkar, Lev Tankelevitch, Ian Drosos, Sean Rintel, Richard Banks, and Nicholas Wilson from Microsoft Research. It was presented at the CHI Conference on Human Factors in Computing Systems in 2025.

As part of the survey, the researchers measured task factors, user factors, and cognitive activities. Under task factors, they studied uses like creation, information, and advice with respect to confidence in their own skills and trust in AI. User factors included a tendency to reflect on work, trust in GenAI, and demographic details. Cognitive abilities study six types of critical thinking—knowledge, comprehension, application, analysis, synthesis, and evaluation.

Those using AI tools tend to think of copy-pasting with minor changes as critical thought, accepting AI responses without any scrutiny in the process. The report has also underscored the relationship between confidence levels and critical thinking. According to the report, higher trust in AI is associated with reduced effort in critical thinking. On the other hand, greater self-confidence results in enhanced critical thinking and in-depth evaluation of responses by AI.

The report states that 60 per cent of participants reported engaging in critical thinking while using Gen AI. Since confidence in Gen AI has been linked to less critical thinking, this means more users trust AI tools and seldom question the outputs. 

What does this mean?

While Gen AI tools boost productivity, they also lead to decreased critical engagement. In low-stakes situations, workers may use AI for draughting emails and social media posts with minor modifications. This leads to the risk of overreliance, and workers may or may not use their independent problem-solving skills. The report is a warning that those who use AI tools extensively may risk cognitive skill erosion, especially in tasks that need deeper analysis.

If you are someone who uses popular AI tools on the work front, you can strike a balance by maintaining an active and questioning mindset, cross-referencing AI responses, and, most importantly, using AI as a copilot and not an ultimate solution provider.