People don’t just distrust ChatGPT. They judge the person who uses it

Imagine this situation: A university student needs a list of books and articles to read for an essay about early forms of art. In one version of the story, he asks a professor for help. In another version, he asks ChatGPT. After that, he plans to collect people's opinions through short surveys. Nobody will be paid to take part. Strangers are asked how many surveys they would be willing to complete.

People were more willing to take part when they were told the student's reading list came from a professor. Everything else was exactly the same: the student, his essay and what he was asking people to do. But when the list came from ChatGPT, people were less willing to help. On average, the student lost about one completed survey because he used AI instead of a professor.

This experiment was one of eight studies in a research paper by Jianning Dang and Li Liu, published in the Journal of Personality and Social Psychology. For years, psychologists have studied something called “AI aversion.” This means people often trust advice from a computer program less than advice from a human, even when the computer gives answers that are just as good or better. Dang and Liu wanted to find out something else: Do people also think less of someone who uses AI? Their research involved 2,317 people in the UK and China. The answer was yes.

What the researchers measured

The researchers used almost the same method in all eight studies. People read a short story about someone asking either an AI tool or a human expert for advice. The stories included a hiring manager reviewing video interviews, someone choosing a wedding gift for a friend, a student, a patient and a primary school teacher preparing reading materials. After reading the story, participants rated how human they thought the person seemed, using two groups of personality traits. The first group was called “human nature.” These included being friendly, enjoying time with other people and trusting others. They are qualities that make humans different from machines. The second group was called “human uniqueness.” These included thinking logically and acting in a mature way. They are qualities that make humans different from animals.

In every study, people who asked AI for advice instead of a human expert were seen as less human. For example, in the wedding gift study, someone who asked ChatGPT for gift ideas received a lower score for seeming human. In most of the studies, the difference was bigger for emotional qualities, such as warmth and friendliness, than for qualities related to logical thinking.

The first study showed this difference especially clearly. It involved 200 people in the UK. They were asked to describe a hiring manager named James in their own words, including what he thought, how he felt and what he wanted to do. People who read that James had used a computer program to help with hiring used fewer words such as “feel” and “care” when describing him. But words such as “plan” and “think” appeared just as often in both groups. This suggests that people still believed the AI user could think and make plans. However, they were less likely to see him as someone with feelings and emotions.

Participants also liked the person who used AI less. But even after the researchers took this difference in liking into account, AI users were still seen as less human. The same result appeared whether participants knew a lot about AI or had very little experience with it. This means that simply being unfamiliar with AI probably does not explain the results.

Who judges AI users more negatively

The researchers believe one reason for these judgments is how similar people think someone is to themselves. Participants felt that people who asked AI for advice were less like them. The researchers' calculations showed that this feeling of difference explained part of why AI users were seen as less human. The negative judgment mainly came from participants who already had negative opinions about AI. Among people who had very positive opinions about AI, the negative judgment mostly went away.

The final two studies looked at whether people would judge AI users less harshly if the AI seemed more like a human. When ChatGPT was described as doing well on tests that measured its understanding of emotions, people judged its users less negatively than when ChatGPT was described as doing poorly on those tests. Even so, people still gave the highest ratings to someone who asked another human for advice instead of a machine.

In other words, making AI seem more human reduced the difference in how people judged AI users by about half. The researchers say this is something companies and designers should think carefully about when trying to make their AI tools seem more alive. They warn that this approach can also have the opposite effect. It can lead to something called the “uncanny valley,” where something that looks or acts almost human makes people feel uncomfortable.

What I think we can learn from this

The finding I would pay the most attention to is what people judge AI users for. They mainly see them as less warm and less able to feel emotions, rather than less capable or less able to think logically. I think using AI has the biggest social cost when people expect you to show real feelings. Choosing a wedding gift and making a hiring decision were both included in the studies. Giving someone feedback about their work or writing a message to an employee who has had a difficult year would be similar situations. These are tasks where people expect another human being to care about how someone feels.

Most discussions about using AI at work focus on whether the tools give answers that are accurate enough to trust. This research raises another question: How do other people see someone who uses AI, and who is making that judgment? Imagine a company that requires employees to use AI, even though many workers still have doubts about the technology. Employees who start using AI early may then seem colder and less caring to their colleagues. The researchers point out that people who feel judged in this way might stop using AI tools. That could hurt the company's efforts to get more employees to use AI.

So what should company leaders do? It would be useful to know which employees will be using AI in situations where feelings matter, especially when other people can see that they are using it. Leaders should also understand how the people receiving that work feel about AI. Not every task should be given to AI. This research warns that automatically turning tasks over to AI can have social costs, even when the technology can do the work. – ©2026 The New York Times Company