Truace
Subscribe
68/100

Moderate evidence

Cite
More
Export CSVExport JSONEmailLinkedInX

Workers engaged more when AI gave criticism and bosses gave praise, experiments with 736 people find

Study title: Which is better? Matching effect on feedback content and source: Gen AI vs. human leaders in driving work engagement

Evidence type
Controlled or validated comparison
Sample
n = 736
Sources
1 sourcePeer-reviewed
Published
2 min read
Truvace record
TRV-2026-2577cite or check it

Study title: Which is better? Matching effect on feedback content and source: Gen AI vs. human leaders in driving work engagement

Three scenario-based experiments with a pooled 736 valid responses tested how the source of performance feedback (generative AI or a human leader) and its content (positive or negative) combine to shape employee work engagement. The matching effect was significant: Gen AI delivering negative feedback and human leaders delivering positive feedback produced the strongest engagement.

Read the study abstract

Purpose Generative artificial intelligence (Gen AI; e.g. ChatGPT, DeepSeek) is increasingly being integrated into performance feedback systems in service organizations, operating alongside human leaders as a dual-source feedback mechanism. Yet the effects of Gen AI feedback on employee work engagement remain unclear. This study addresses the central question of whether and how the matching between feedback source (Gen AI vs. human leaders) and feedback content (positive vs. negative) differentially influences em…

Which is better? Matching effect on feedback content and source: Gen AI vs. human leaders in driving work engagement
"Scottish Parliament Debating Chamber 3" by Colin, CC BY-SA 4.0.

The gain

In three scenario experiments with 736 valid responses, employees showed higher work engagement when Gen AI delivered negative feedback and human leaders delivered positive feedback.

In brief

The finding matters for service organizations that are adding Gen AI, such as ChatGPT or DeepSeek, to feedback systems that run alongside human managers. The results suggest the two sources have complementary strengths rather than one simply replacing the other.

The researchers traced the effect to two kinds of motivation, promotion focus and prevention focus, and found that self-efficacy changed how strong the effect was. All three studies used scenarios rather than live workplaces, so they measure responses to described situations.

Main points

  1. The research combined three scenario-based experiments with 736 valid responses in total.
  2. It tested feedback source (Gen AI or human leader) against feedback content (positive or negative).
  3. Gen AI plus negative feedback, and human leaders plus positive feedback, were the pairings that best raised work engagement.
  4. Promotion-focused and prevention-focused regulation mediated the effect.
  5. Employee self-efficacy moderated the effect.
  6. The authors build on signaling theory and regulatory focus theory and cite the dehumanized nature of Gen AI as part of the explanation.

The full story

The study responds to the spread of Gen AI in performance feedback in service organizations, where it increasingly works alongside human leaders as a second source of feedback. The authors say the effect of that AI feedback on employee work engagement had been unclear.

Using three scenario-based experiments, they compared who gives the feedback with whether it is positive or negative. The best results came from complementary pairings: AI for negative feedback and human leaders for positive feedback.

Further analyses showed that promotion-focused and prevention-focused regulation explained the effect, and that self-efficacy moderated it. The authors offer practical guidance for designing dual-source feedback systems that optimize employee engagement.

Sources

1 on file
  1. Journal of Service Theory and Practicedoi:10.1108/jstp-11-2025-0452
    Peer-reviewed2026-06-07

Talk it over

Four questions to start onWho gets this benefit, and who was left out of the study or story?

Start with a question

Workers engaged more when AI gave criticism and bosses gave praise, experiments with 736 people findhttps://truvace.com/item/which-is-better-matching-effect-on-feedback-content-and-source-gen-ai-vs-human-l

  1. Who gets this benefit, and who was left out of the study or story?

    The gain In three scenario experiments with 736 valid responses, employees showed higher work engagement when Gen AI delivered negative feedback and human leaders delivered positive feedback.

    Answer in the thread

  2. How 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 736, from Journal of Service Theory and Practice.

    Answer in the thread

  3. What does it leave open? What else would you want to know before acting on it?

    The record’s own caveat The evidence comes from scenario-based experiments rather than observed workplace feedback, so it reflects responses to described situations.

    Answer in the thread

  4. Have 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

More from Labor