The appeal of using artificial intelligence (AI) on the job is obvious. AI can produce faster drafts, cleaner summaries, even instant answers. But new research suggests there may be a subtle trade-off that has less to do with artificial intelligence itself and more to do with how we feel about our own ability to think.

A new study finds that relying heavily on AI tools at work can chip away at confidence in our own intelligence and ability to reason and our sense of ownership over ideas.

The research looked at over 1,900 adults in the United States and Canada.

Letting AI do your thinking for you can make you unsure of your ability to figure things out.

People in the study were asked to use commercially available AI tools to complete 10 simulated workplace tasks. These included developing plans with incomplete information, interpreting ambiguous data and explaining strategic decisions — the kinds of thinking many so-called “knowledge workers” do every day.

What the researchers were interested in finding out wasn't whether AI helped people perform better, but how people interacted with it and how that interaction shaped their confidence.

After completing the tasks, 58 percent of participants said AI had done most of the thinking for them. This perception mattered. Those individuals reported lower confidence in their own independent reasoning and a weaker sense that the final ideas were truly theirs.

Study participants also described making a trade-off: gaining speed and efficiency but losing depth of thought. “When we look at brain activity contingent on how people choose to use the tool, we can see increases or decreases. It really doesn't have to do with the tool itself,” Baldeo explained in a press release.

Participants who took a more active role, such as editing AI outputs, questioning them or even rejecting them, reported something very different.

The effect was especially pronounced in tasks involving planning and sequencing. These are areas where AI excels and where it's easy to hand over the cognitive heavy lifting. But the study's findings weren't one-sided.

Participants who took a more active role, such as editing AI outputs, questioning them or even rejecting them, reported something very different. They said that they experienced greater confidence and a stronger sense of authorship.

The distinction between using AI as a collaborator versus as a crutch is an important finding. “The issue was not AI use itself but the degree of passive acceptance,” lead author Sarah Baldeo, MBA, a PhD candidate in AI and neuroscience at Middlesex University in the United Kingdom, said.

The study also revealed some interesting patterns. For instance, men reported higher levels of reliance on AI than women. And more experienced participants were more likely to challenge AI than women. Experienced participants were also more likely to challenge AI outputs, suggesting that expertise may act as buffer.

It's important to note that the study is correlational; it shows that there is a relationship between using AI and diminishing confidence in one's ability to think, but it doesn't prove that AI use directly causes reduced confidence in our cognitive abilities. Still, the pattern is consistent enough to raise questions about how these tools are shaping not just what we produce, but how we experience our own thinking.

This research taps into a growing concern about “cognitive offloading” which is the tendency to outsource mental effort to technology. We've seen it before with calculators, GPS and search engines. But AI is different in one key takeaway: it doesn't just retrieve information, it generates ideas. That can blur the line between assistance and authorship.

AI doesn't inherently make us less capable. But it can change how we relate to our own abilities.

As Baldeo puts it: “Broadly, the best way to use AI is to train it rather than letting it train you.” The study offers several suggestions about how to do that:

  • Try solving a problem yourself before turning to AI.
  • When you do use AI, refine your prompts multiple times instead of accepting the first one.
  • Consider taking regular breaks from using AI tools to keep your own cognitive abilities sharp.

Perhaps the most important insight from the study is this: AI doesn't inherently make us less capable. But it can change how we relate to our own abilities. It turns out, the challenge isn't whether to use it, but rather how to use it without stepping out of the thinking process altogether.

The study is published in the American Psychological Association's Technology, Mind, and Behavior.