Is Smarter AI Making People Think Less?
Short answer: Research by Carnegie Mellon and Microsoft Research, surveying 319 knowledge workers, found that people who trust AI too much use less critical thinking, while people with high self-confidence keep thinking critically even with AI help. In 55–79% of the examples respondents gave, depending on the activity, they felt AI-assisted work took less or much less critical-thinking effort, most often for comprehension and least often for evaluating quality; these are self-reported perceptions, not measurements. People who do think critically with AI set clear goals, verify information, and adapt the output to their real context.
This post is based on research by Carnegie Mellon and Microsoft Research that surveyed 319 knowledge workers who use AI (such as analysts, lawyers, engineers, doctors, and developers).
Generative AI (such as ChatGPT and Copilot) is changing people's ability to think critically.
- —People who trust AI too much = use less critical thinking
- —People with high self-confidence = keep thinking critically even with AI help
- —Easy, low-risk tasks = people often check nothing at all and let AI do the work
- —Hard, high-risk tasks = people still try to check information, correct it, and adapt it
Is AI making us lazy thinkers?
The research team found that, depending on the type of activity, respondents said they used "less" or "much less" critical-thinking effort with AI in 55–79% of the examples they described: highest for comprehension (79%) and lowest for evaluating quality (55%). These figures reflect respondents' own perceptions, not a direct measurement of effort.
The critical thinking that remains
When people use AI and genuinely think critically, they usually do three things:
- —1) Set clear goals: know what they want from the AI (rather than typing something random and hoping for a good result)
- —2) Verify information: check the sources the AI cites and cross-check against outside information
- —3) Adapt or add to the output: take the AI's results and tailor them to the real context
Why do some people do no critical thinking at all when using AI?
The analysis sorted the answers into three broad groups:
- —Lack of awareness: they don't know what to watch out for, and simply trust the AI
- —Lack of motivation: they have no incentive, because "the task isn't important" or "time is limited"
- —Lack of ability: they don't know enough to tell when the AI is wrong, or can't fix it themselves
In short
AI may help us work faster, but it also risks quietly weakening our critical-thinking skills, especially if we use it without asking questions.
Right now, a clear split is emerging among working people:
- —People who "use AI as an assistant" and still think critically
- —People who have "become dependent on AI" without realizing it
"Who is controlling whom" may be the most important question of this era.
Read the full research paper.
Translated from the Thai original by NXT Consulting Group.
Share this article
