.
A survey of hundreds of professionals has revealed that high confidence in generative AI tools like ChatGPT is linked to a decrease in critical thinking.
Lead author Hao-Ping Lee of Carnegie Mellon University, along with a team from Microsoft Research, detailed these findings in the study “The Impact of Generative AI on Critical Thinking: Self-Reported Reductions in Cognitive Effort and Confidence Effects From a Survey of Knowledge Workers.”
The study suggests that while AI can handle routine work, it may lead to “cognitive offloading,” where users stop scrutinizing the technology’s output as they become more comfortable with its performance.
Love reading and writing? Submit a post to our blog.
The authors write: “Specifically, higher confidence in GenAI is associated with less critical thinking, while higher self-confidence is associated with more critical thinking.”
To reach these conclusions, researchers surveyed 319 knowledge workers across various industries, including technology, education, and finance.
Participants shared 936 real-world examples of how they used generative AI for tasks like writing code, drafting emails, or brainstorming project ideas.
Using statistical models to analyze these examples, the researchers measured how “task-specific self-confidence” and “confidence in GenAI” influenced whether a worker actually stopped to think critically about the AI’s response.
The data showed a significant correlation where users who viewed the AI as highly competent were less likely to put in the mental effort to verify or challenge its output.
The study also found that generative AI is fundamentally changing what it means to think critically on the job, shifting the focus away from creating ideas and toward overseeing, verifying, and integrating AI output into their final products.
“Qualitatively, GenAI shifts the nature of critical thinking toward information verification, response integration, and task stewardship,” the researchers noted.
Another finding identified that the primary motivators for critical thinking were the desire for high-quality work and the fear of negative consequences. Many participants reported that they engaged their “critical thinking muscle” when they felt a task was high-stakes or when they were worried about a biased or incorrect output from AI.
Furthermore, the researchers identified the lack of “awareness” as a major barrier to critical thinking. In many cases, users didn’t even realize that a task required critical thought because the AI produced a polished and authoritative response, leading to a false sense of security.
The paper acknowledges several limitations, noting that the results are based on the self-reported perceptions of workers rather than direct behavioral measurements. Because participants had to recall and interpret their own thinking processes, there is a risk that they may have conflated the general ease of using AI with a genuine reduction in their critical thinking effort.
Looking ahead, the researchers call for the development of AI tools that include “cognitive forcing functions” — design features that intentionally slow the user down or prompt them to double-check specific parts of an AI’s answer.
I’m a freelance writer with 6 years of experience in SEO blogging and article publishing. While you’re here, get the latest updates by subscribing to my newsletter.
