⚡
Brief IA
›

77 Adults: Unlabeled AI Images Perceived as Well as Stock Photos

🛠️ AI Tools·Tom Levy·

77 Adults: Unlabeled AI Images Perceived as Well as Stock Photos

77 Adults: Unlabeled AI Images Perceived as Well as Stock Photos
⚡
Key Takeaways
177 adults evaluated 6 web pages, including 3 AI-generated images and 3 stock photos
2Authors: no more negative perception for unmarked AI images
3Analyses: mixed effects, paired t-tests, and Bonferroni correction
💡Why it matters — without origin mention, the AI-generated images tested did not degrade judgments compared to stock photos.
⚡Le brief IA que lisent les pros

Le brief IA que les pros lisent chaque soir

Les 7 actus IA du jour, décryptées en 5 min. Gratuit.

Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.

Choisis ton rythme

Gratuit · Pas de spam · Désabonnement en 1 clic

A comparison was conducted among 77 adults in the United States between identical web pages where only the main image varied between real stock photos and AI-generated visuals. After 10 seconds of exposure, participants rated the reliability, professionalism, and authenticity of the fictitious company. In the absence of information about the source of the images, the tested AI visuals did not lead to a more negative perception than stock photos.

77 adults evaluate 6 pages; mixed-effects analyses and t-tests

The authors of the study recruited 77 adults from the general population of the United States. Each participant evaluated 6 versions of a fictitious consulting company’s webpage, identical except for the main image. Three versions used AI-generated stock images, while three used real stock photos. Participants were not informed that AI-generated images were involved. After 10 seconds of viewing each page, they rated the reliability, professionalism, and authenticity of the company, then answered an open-ended question about other factors that influenced their impression. For the quantitative analysis, the authors employed a mixed-effects model to account for differences between participants and images, supplemented by paired t-tests image by image with Bonferroni correction for multiple comparisons. The open-ended responses were subjected to thematic analysis.

Without mention of AI, no perceived decline compared to stock photos

The stated objective was to assess whether AI-generated images, encountered without source information, alter judgments about a company. Under these conditions, the tested AI images did not lead to a more negative perception than stock photos. They argue more broadly that AI-generated visuals can compete with stock photography. They also note that some organizations remain hesitant, fearing that their image may be perceived as less reliable, less professional, or less authentic than that conveyed by real photographs. Furthermore, they specify that AI tools allow for the rapid and cost-effective generation of visuals for websites, which is becoming an increasingly common solution for design teams.

⚡

Brief IA — L'actualité IA en français

L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.