aceModerate evidence
More
Praise from a more human-like AI led workers to take more advice, experiment finds
Study title: AI ‐Enabled Performance Feedback Design: The Effects of Anthropomorphic AI Feedback Source, Feedback Valence and Emotional Arousal on Employee Advice‐Taking Behaviour
- Evidence type
- Controlled or validated comparison
- Sources
- 1 sourcePeer-reviewed
- Published
- 2 min read
- Truvace record
- TRV-2026-2696cite or check it
Study title: AI ‐Enabled Performance Feedback Design: The Effects of Anthropomorphic AI Feedback Source, Feedback Valence and Emotional Arousal on Employee Advice‐Taking Behaviour
Researchers ran a 2×2 experiment on how employees respond when an AI system gives them performance feedback. They varied how human-like the AI feedback source was and whether the feedback was positive or negative, then measured advice-taking in a production task.
Read the study abstract
ABSTRACT Companies are increasingly using artificial intelligence (AI) to provide employees with performance feedback, but its efficacy is inconsistent. Drawing upon the theories of mind perception and feedback intervention, we examine the effects of anthropomorphic AI feedback source and feedback valence in conveying performance feedback on employees' advice‐taking behaviours. We conducted a 2 × 2 experiment in which we manipulated anthropomorphic AI feedback source (more versus less) and examined how feedback…
The gain
In a 2×2 experiment using skin-conductance sensors during a production task, employees who got positive feedback from a more human-like AI showed higher emotional arousal and took more advice afterward than those who got it from a less human-like AI.
In brief
Positive feedback from the more human-like AI produced higher emotional arousal, recorded with an electrodermal activity device, and more advice-taking afterward than positive feedback from the less human-like AI.
With negative feedback, arousal and advice-taking did not differ between the more and less human-like AI. The result points to emotional language mattering most when feedback is favourable, as companies roll out AI feedback tools with uneven results.
Main points
- The experiment crossed two factors: a more versus less human-like AI feedback source and positive versus negative feedback.
- Emotional arousal was measured with an electrodermal activity device while employees did a production task.
- Positive feedback from a more human-like AI led to higher arousal and more advice-taking than from a less human-like AI.
- With negative feedback, arousal and advice-taking were the same for both AI types.
- The authors draw on mind perception and feedback intervention theories to explain the pattern.
The full story
The study starts from the observation that companies increasingly use AI to give employees performance feedback, but that its effectiveness is inconsistent. It tests whether making the AI source seem more human changes how employees react.
Participants did a production task and received feedback from either a more or a less anthropomorphic AI. Skin-conductance readings captured emotional arousal, and later advice-taking behaviour was the outcome of interest.
The authors report that the human-like source only made a difference for positive feedback, where it raised arousal and advice-taking. They frame this as evidence for the role of emotional language in feedback delivered by anthropomorphic AI systems.
Sources
1 on file- Accounting & Financedoi:10.1111/acfi.70228
Talk it over
Who gets this benefit, and who was left out of the study or story?
Praise from a more human-like AI led workers to take more advice, experiment findshttps://truvace.com/item/ai-enabled-performance-feedback-design-the-effects-of-anthropomorphic-ai-feedbac
Who gets this benefit, and who was left out of the study or story?
The gain In a 2×2 experiment using skin-conductance sensors during a production task, employees who got positive feedback from a more human-like AI showed higher emotional arousal and took more advice afterward than those who got it from a less human-like AI.
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, from Accounting & Finance.
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 human-like AI advantage appeared only for positive feedback; under negative feedback, arousal and advice-taking did not differ between the AI types.
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