aceModerate evidence
More
Humanlike AI boosted work engagement in 501-person experiments, until workers felt threatened
Study title: Beyond Human–AI interaction: exploring the influence of AI anthropomorphism on work engagement
- Evidence type
- Controlled or validated comparison
- Sample
- n = 501
- Sources
- 1 sourcePeer-reviewed
- Published
- 2 min read
- Truvace record
- TRV-2026-2501cite or check it
Study title: Beyond Human–AI interaction: exploring the influence of AI anthropomorphism on work engagement
A peer-reviewed study built from two pilot studies and two main experiments, 501 participants in total, tested how humanlike cues in AI affect employee work engagement. Participants worked through immersive workplace role-playing scenarios in which collaboration with AI was mandatory and performance-oriented.
Read the study abstract
Purpose This study examines how anthropomorphic cues that make AI agency socially salient influence employee work engagement in mandatory, performance-oriented human–AI collaboration. It distinguishes AI agency from AI anthropomorphism, tests affective and evaluative motivational mechanisms and identifies perceived AI threat as a boundary condition grounded in the Job Demands-Resources (JD-R) model. Design/methodology/approach Across two pilot studies and two main experiments (total N = 501), we used immersive w…
54Gain side
How the two sides scoreClose call
58Problem side
The gain
In experiments with 501 participants, anthropomorphic AI raised work engagement through emotional bonding with the AI and through a motive to make a good impression on it.
The problem
In the same experiments, when participants perceived a high AI threat, the positive effects of anthropomorphism on emotional bonding, impression management motive and work engagement became non-significant.
A fixed formula sets both scores from the evidence and its sources. Reader votes don’t move them.
What the scores mean
G is the gain side, out of 100. It goes up when the gain is specific, when it measurably helped people and when the sources behind it are strong.
P is the problem side, out of 100. It goes up when the problem is specific, when it measurably hurt people and when the sources behind it are strong.
In brief
Anthropomorphic AI raised engagement by two routes: an emotional one, in which people bonded with the AI, and an evaluative one, in which they wanted to make a good impression on it. The authors call the second route performing for AI.
The benefit depended on how threatened people felt. When perceived AI threat was high, the effects of humanlike design on bonding, impression management and engagement were no longer significant. The authors suggest replacement-related threat drains the psychological capacity that humanlike cues need to work as job resources.
Main points
- The study combined two pilot studies and two main experiments with a total of 501 participants.
- Scenarios were immersive workplace role-plays of mandatory, performance-oriented human-AI collaboration.
- Emotional bonding with the AI was one route linking humanlike design to higher work engagement.
- Impression management motive, or performing for the AI, was a second route to higher engagement.
- High perceived AI threat made the effects on bonding, impression management and engagement non-significant.
- The authors frame threat as a hindrance demand within the Job Demands-Resources model.
- The findings offer a contingency explanation for the anthropomorphism paradox: benefits of humanlike design depend on the surrounding AI-threat climate.
The full story
The research separates AI agency from AI anthropomorphism. It examines humanlike cues that make an AI's agency socially salient, without treating anthropomorphism as the same thing as agency.
The team combined social response theory with the Job Demands-Resources model. Their dual-pathway model has anthropomorphic design acting on engagement through emotional bonding and through an acquisitive impression management motive, with perceived AI threat acting as a hindrance demand that weakens both paths.
Results supported both paths. Humanlike AI raised engagement through bonding and through the wish to perform well for the AI, but the effects were significant only when perceived threat was not high.
For workplaces, the findings suggest that making AI more humanlike may not raise engagement where employees fear being replaced. The authors say the payoff of humanlike design depends on the AI-threat climate around it.
What this doesn’t fix
The findings come from role-playing scenarios in experiments rather than from observation of employees in real workplaces.
Sources
1 on file- Information Technology & Peopledoi:10.1108/itp-06-2025-0862
Talk it over
This record found a gain and a problem in the same place. Which matters more to you, and who might answer differently?
Humanlike AI boosted work engagement in 501-person experiments, until workers felt threatenedhttps://truvace.com/item/beyond-human-ai-interaction-exploring-the-influence-of-ai-anthropomorphism-on-wo
This record found a gain and a problem in the same place. Which matters more to you, and who might answer differently?
The gain In experiments with 501 participants, anthropomorphic AI raised work engagement through emotional bonding with the AI and through a motive to make a good impression on it.
The problem In the same experiments, when participants perceived a high AI threat, the positive effects of anthropomorphism on emotional bonding, impression management motive and work engagement became non-significant.
Answer in the threadHow do we know? What can this kind of evidence show, and what can’t it?
The evidence This rests on a controlled comparison with a sample of 501, from Information Technology & People.
Answer in the threadWhat does it leave open? What else would you want to know before acting on it?
The record’s own caveat The findings come from role-playing scenarios in experiments rather than from observation of employees in real workplaces.
Answer in the threadHave you seen AI used this way where you study, work or live? Did it go the same way?
Answer in the thread
The questions are the same on every record. The details under them are taken from this record as written.
What readers say