Effect of AI authorship disclosure on purchase intention in cultural tourism souvenirs: The mediating roles of perceived effort and authenticity

As generative artificial intelligence (AI) expands in cultural tourism souvenir design and disclosure policies evolve, evidence remains limited on how AI authorship disclosure shapes responses across nationality groups and products. Drawing on signaling theory, two studies compared consumers from China and the U.S. across three disclosure conditions: unlabeled, human–AI co-created, and AI-generated. Study 1 used tourism postcard stimuli ( N = 1110), and Study 2 used mug and tourism tote bag stimuli ( N = 1255).…

Effect of AI authorship disclosure on purchase intention in cultural tourism souvenirs: The mediating roles of perceived effort and authenticity
هتل داريوش نماد ميهمان نوازي پارسيHotel hospitality symbol of Dariush the Persian - panoramio (1) by alirezaseydabady. CC BY 3.0 · https://creativecommons.org/licenses/by/3.0

In brief

Two experiments compared consumers from China and the U.S. who saw cultural tourism souvenirs labeled in one of three ways: unlabeled, human-AI co-created, or AI-generated. Study 1 used postcards and Study 2 used mugs and tote bags, with more than 2,300 participants in total.

The results suggest that revealing AI involvement can cost sellers sales, and that the effect depends on market and product. It remains unclear how far the findings extend beyond the products and two nationalities tested, and the human-AI co-creation label did not reliably help.

Main points

  1. Disclosure affected purchase intention and interacted significantly with nationality in both studies.
  2. AI-generated labels consistently produced lower purchase intention than no label.
  3. Human-AI co-created labels performed differently across the two studies.
  4. Perceived effort and perceived authenticity mediated the disclosure effects.
  5. Nationality moderated some, but not all, of the indirect effects.

The problem

Telling consumers that souvenirs are AI-generated reduced purchase intention relative to no disclosure, with the drop significant among U.S. participants but not Chinese participants.

The rundown

The study drew on signaling theory and tested three disclosure conditions. Study 1 (N = 1110) used tourism postcards, while Study 2 (N = 1255) used mug and tourism tote bag stimuli.

In Study 1, unlabeled and human-AI co-created conditions did not differ and both exceeded the AI-generated condition. In Study 2, the unlabeled condition exceeded both disclosed conditions for both products, and the two disclosed conditions did not differ.

Parallel mediation analyses found indirect effects through perceived effort and perceived authenticity. The human-AI co-created versus unlabeled authenticity effect was nonsignificant in Study 1 but significant in Study 2. The authors advise firms not to assume co-creation helps, to clarify the human contribution, and to pretest disclosure strategies by product and market.

Sources

  1. Peer-reviewedJournal of Retailing and Consumer Services2026-10-20

The debate