Study finds ChatGPT users are judged as less human by observers

People who use ChatGPT for everyday tasks face a social penalty, with observers judging them as colder and less human, according to psychology research published in the Journal of Personality and Social Psychology by Jianning Dang and Li Liu.

The Tech TL;DR:

• People who rely on AI tools like ChatGPT instead of human experts face a distinct social penalty, being perceived as having lower emotional capacity.
• Across eight studies involving 2,317 participants in the UK and China, researchers found the humanness gap concentrates heavily on warmth and emotional traits rather than rational competence.
• This social friction can impact enterprise software adoption, as employees who visibly use AI for emotionally loaded tasks risk being viewed as colder by their colleagues.

University Students and Wedding Gift Experiments Reveal AI Stigma

The research design tested how everyday participants evaluate individuals seeking advice from algorithms versus human experts. In one scenario, a university student needed a reading list for an essay on primitive art. When told the reading list came from a professor, strangers were more likely to sign up for unpaid feedback surveys compared to when the identical list came from ChatGPT. The student lost roughly one survey’s worth of goodwill simply because of the tool he used.

Across eight separate studies, Jianning Dang and Li Liu evaluated 2,317 participants in the UK and China. Test subjects read scenarios covering a hiring manager screening video interviews, a person choosing a wedding gift, a student, a patient, and a primary school teacher preparing reading materials. Participants then rated the individuals on traits related to dehumanization, separating human nature traits like friendliness and sociability from human uniqueness traits like rationality and maturity.

Dehumanization Concentrates on Emotional Warmth Over Competence

When individuals consulted AI instead of a human expert, observers dehumanized them more. In the wedding gift experiment, the person who asked ChatGPT for ideas scored lower on the humanness scale. The deficit appeared most sharply on emotional traits rather than rational ones. In the first study involving 200 UK participants who wrote free-form descriptions of a hiring manager named James, respondents who learned James used a recruiting algorithm used fewer words like feel and care. Words like plan and think appeared at an equal rate in both groups.

Respondents granted the AI user a working mind equipped with rational capabilities, but expressed uncertainty regarding emotional states. Participants liked the AI user less overall, though the humanness gap persisted even after controlling for personal preference. The effect remained constant regardless of participants' prior experience or familiarity with artificial intelligence.

Perceived Similarity Influences Social Penalties for AI Users

The underlying psychological mechanism points toward perceived similarity. Participants viewed the AI advice seeker as less like themselves, a social distance that statistically accounted for part of the humanness gap. This effect ran primarily through individuals who held negative views of technology, while participants with highly favorable views of AI showed virtually no negative judgment.

The final two studies tested whether the social penalty shrank when the AI displayed humanlike traits. When ChatGPT scored well on tests of emotional understanding, the user faced a smaller penalty compared to instances where the AI scored poorly. Anthropomorphizing the tool closed roughly half the penalty gap, though the highest ratings still went to individuals who consulted humans instead of machines.

AI Adoption Creates Social Friction in Emotional Tasks

These findings suggest that AI adoption costs the most in tasks requiring emotional resonance. Wedding gifts, hiring decisions, performance feedback, and employee messages fall into categories where observers expect a human to feel something. Organizations mandating AI tools while a portion of the workforce remains skeptical risk exposing early adopters to social friction from colleagues who perceive them as colder.

Product designers face a delicate balance when attempting to anthropomorphize AI tools, as excessive human-mimicry can backfire into the uncanny valley. Business leaders assessing automation strategies must account for how visibly deploying algorithms for emotionally loaded work alters interpersonal perceptions among staff.

Disclaimer: The technical analyses and security protocols detailed in this article are for informational purposes only. Always consult with certified IT and cybersecurity professionals before altering enterprise networks or handling sensitive data.