Truace
Subscribe
68/100

Moderate evidence

Cite
More
Export CSVExport JSONEmailLinkedInX

In a 346-shopper experiment, objectivity built trust only when a chatbot gave green claims

Study title: Luxury fashion brands’ AI-powered sustainability chatbot communication: how information qualities and brand luxury perception shape consumer trust and behavior

Evidence type
Controlled or validated comparison
Sample
n = 346
Sources
1 sourcePeer-reviewed
Published
3 min read
Truvace record
TRV-2026-2644cite or check it

Study title: Luxury fashion brands’ AI-powered sustainability chatbot communication: how information qualities and brand luxury perception shape consumer trust and behavior

Researchers at Seoul National University and Kunsan National University showed 346 US consumers aged 20 to 49 a chat exchange about a sustainable Gucci handbag. The sustainability claims were identical, but the provider, named Alex, appeared either as a human service representative (171 people) or as a robot chatbot (175 people).

Read the study abstract

This study examines how sustainability communication in luxury retail shapes brand trust and behavioral intentions, focusing on consumers' post-exposure perceptions of information quality and provider type. Building on perceived objectivity and perceived persuasiveness, the study explores how consumers evaluate identical sustainability information attributed to either a human agent or an artificial intelligence (AI) chatbot, and how these perceptions influence brand trust, information acceptance intention, and p…

The gain

In a scenario experiment with 346 US online consumers shown identical Gucci sustainability messages, perceived objectivity raised purchase intention through brand trust only when the messages came from an AI chatbot, not from a human agent.

In brief

Perceived objectivity of the information fed into brand trust, and then into purchase intention, only in the chatbot condition. The indirect effect on purchase intention was .379 for the chatbot group versus .055 for the human group, whose confidence interval included zero. Perceived persuasiveness worked equally well under both providers.

The chatbot advantage also depended on how luxurious people found the brand. Among those who rated Gucci lower on luxury, trust translated into purchase intention more strongly under the chatbot, while higher-luxury consumers showed no significant provider difference. The authors caution that this is subgroup evidence, not a formal interaction test, and that all outcomes were stated intentions in an imagined scenario.

Main points

  1. Participants were randomly assigned to a human-agent (171) or AI-chatbot (175) version of the same chat; message content, tone and structure were identical, with only the provider image changing.
  2. Perceived objectivity (β=.291) and perceived persuasiveness (β=.392) both increased brand trust, and trust increased information acceptance (β=.552) and purchase intention (β=.570).
  3. The objectivity-to-trust-to-purchase-intention effect was .379 under the chatbot and .055 (not significant) under the human agent; the difference was significant (p<.001).
  4. The persuasiveness path was significant and similar for both providers (for information acceptance, .214 human versus .212 chatbot), so chatbots gave no extra boost there.
  5. Among 165 lower-luxury-perception consumers, the trust-to-purchase-intention link was .715 under the chatbot versus .377 under the human agent (p=.020).
  6. Among 181 higher-luxury-perception consumers, no significant provider differences appeared on any path.
  7. The stimulus used 2025 Gucci and Kering targets, such as cutting environmental impact by 50% and regenerating one million hectares of land, phrased as progress already made.

The full story

The experiment recruited 360 US consumers in their 20s to 40s through Amazon Mechanical Turk, and 346 remained after condition-recognition and attention checks. Participants studied the scenario images for at least three minutes, then rated perceived objectivity, persuasiveness, brand trust, information acceptance, purchase intention and perceived brand luxury on seven-point scales. Pre-existing attitudes toward Gucci did not differ between groups (human mean 6.152, chatbot 6.028).

Data were analysed with partial least squares structural equation modeling and bootstrapped multi-group analysis. Brand trust explained 35.8% of the variance in trust itself's predictors' outcome model (R² .358 for trust, .304 for information acceptance and .326 for purchase intention). Out-of-sample prediction beat a naive benchmark but was low compared with linear regression on most indicators.

Objectivity's trust-mediated effects on information acceptance and purchase intention differed by provider (differences of -.275 and -.323). For the chatbot group, the effects were .334 and .379; for the human group, .059 and .055, both not significant. Persuasiveness effects did not differ by provider.

Perceived brand luxury was split at the mean (5.671) into low (165) and high (181) groups. Under low luxury perception, trust predicted information acceptance and purchase intention more strongly with the chatbot (differences -.242, p=.021 and -.337, p=.020), which the authors read as partial support for a brand-luxury boundary condition.

Sources

1 on file
  1. Fashion and Textilesdoi:10.1186/s40691-026-00493-8
    Peer-reviewed2026-10-23

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

In a 346-shopper experiment, objectivity built trust only when a chatbot gave green claimshttps://truvace.com/item/luxury-fashion-brands-ai-powered-sustainability-chatbot-communication-how-inform

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

    The gain In a scenario experiment with 346 US online consumers shown identical Gucci sustainability messages, perceived objectivity raised purchase intention through brand trust only when the messages came from an AI chatbot, not from a human agent.

    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 346, from Fashion and Textiles.

    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 brand-luxury findings come from split-sample subgroup comparisons rather than a formal interaction test, and two measures (perceived objectivity and purchase intention) had low Cronbach's alpha (.519 and .559), so results should be treated cautiously.

    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 Lifestyle