AI-Generated Images Can Carry out as Well as Inventory Images

AI instruments can now generate polished web site imagery shortly and at low value, making them an more and more sensible choice for design groups. But organizations could hesitate to make use of AI-generated photos in the event that they fear that customers will see them as much less reliable, skilled, or genuine than actual images. We performed a research to check whether or not AI imagery, when encountered with out provenance info, impacts customers’ precise perceptions of an organization.

How We Evaluated Perceptions of Company Trust Between AI vs. Real Images

We recruited 77 individuals from the overall U.S. grownup inhabitants to judge 6 variations of a fictional consulting agency’s webpage. The webpages had been equivalent apart from the hero picture: 3 used AI-generated inventory photos and three used actual inventory photos. Participants weren’t advised that the research concerned AI-generated photos. After viewing every web page for 10 seconds, individuals rated the corporate’s trustworthiness, professionalism, and authenticity. They additionally answered an open-ended query: “What other factors influenced your impression of this company?”For the quantitative evaluation, we performed a mixed-effects evaluation to account for variations throughout individuals and pictures. We additionally ran pairwise t-tests evaluating scores for every particular person picture and utilized a Bonferroni correction to account for a number of comparisons.

We analyzed the open-ended responses utilizing thematic evaluation.

Websites with AI-Generated Images Were Not Inherently Less Trustworthy

We anticipated websites utilizing AI-generated imagery to be perceived much less favorably than these utilizing actual images. Instead, we discovered no proof that the AI-generated photos we examined lowered belief within the firm’s web site. Ratings had been barely greater for webpages with AI-generated photos on all three measures, however just for authenticity did the distinction between AI and inventory photos attain statistical significance. Nevertheless, even for authenticity, the estimated score enhance as a result of AI picture was very small (solely 0.4 factors on a scale from 1 to 7).

All the t-tests confirmed no statistically important variations in trustworthiness, authenticity, or professionalism (except exams involving one picture).

Webpages utilizing each AI-generated photos and actual photos acquired many constructive open-ended feedback indicating that they had been reliable, largely on account of depictions of collaboration and teamwork within the photos.

Regarding AI Image 1:

“I like that they are all working together and that makes me feel like this company values teamwork, and that to be is indicative of a company that is big and reputable.”

Regarding Real Image 2:

“I think the biggest factor that influenced me was the image of the people meeting together and seeming to be in unison. As though they came to a resolution or made a plan, it made me feel at ease about the company, as though they could be trusted.”

One actual picture (Real Image 1) elicited a number of unfavorable feedback about illustration, hierarchy, and physique language.  

Regarding Real Image 1:

“I don’t notice the diversity as much”

“The group does not seem to be diverse enough”

“The individual leaning in doesn’t give me a collaborative feel”

“The image felt like there was a leader and it made me less trusting of this page”

Thus, despite the fact that throughout all 3 measures, web sites with AI-generated photos acquired barely greater scores, as a result of the variations had been small and our research included solely 6 photos, we can not conclude that AI photos are categorically higher than inventory photos. What our outcomes present is that the AI-generated photos weren’t at drawback.

These findings must be interpreted inside the context of this research. We examined a small set of office photos on a fictional consulting-company web site, so the outcomes could not generalize to different sorts of imagery, industries, or contexts.

Cultural and Gender Representation Shaped Participants’ Perceptions

Racial and gender variety had been talked about repeatedly when individuals defined why they appreciated or disliked specific photos.

Several of the AI-generated photos in our research prompted constructive feedback about perceived variety and inclusion. Additionally, one of many better-received actual photos depicted a racially numerous group with a girl in a management position.

Regarding AI Image 3:

“I like this picture more than the others. It shows the company has inclusion and a good teamwork system.”

Regarding Real Image 3:

“This one had more diversity in the workplace, which overall gave me a better feeling.”

This is a crucial inclusive-design consideration. Users typically worth seeing individuals much like themselves, and their perceptions of belief with a corporation will be pushed partly by whether or not its imagery displays variety and inclusion.

If groups choose to make use of AI to generate photos, it’s vital to notice that AI doesn’t routinely produce inclusive imagery. Teams should fastidiously assess whether or not picture candidates uphold inclusive-design ideas. This entails wanting past the presence of variety alone: think about who’s positioned in management or supporting roles, how completely different individuals are positioned and interacting, and whether or not the picture reinforces stereotypes about gender, race, age, capacity, or career.

Perceiving an Image as AI-Generated May Still Affect Reactions

Commenting on whether or not an image was AI-generated was not a high precedence for the overwhelming majority of individuals. However, a couple of individuals shared their suspicions of AI use within the open-ended responses. Sometimes, actual photos had been incorrectly categorized as being AI-generated.

When individuals commented that they believed the imagery was AI-generated, they tended to additionally price the positioning together with it much less favorably.

Regarding Real Image 3:

“The image seems again like its fake/AI generated. The main blonde woman in the photo looks like she could be a real worker but her having the ‘spotlight’ in the image makes it all seem a bit fake.”

This creates an vital distinction for organizations: Users should react negatively once they suspect that AI was concerned, even when an AI-generated picture is thoughtfully created and seems pure. Since individuals in our research weren’t advised that some photos had been AI-generated, these findings don’t inform us how the identical photos would carry out in the event that they had been clearly recognized as AI-generated.

That distinction is turning into more and more related as authorized necessities for AI transparency proceed to evolve. For instance, the European Union’s AI Act contains necessities for figuring out or disclosing sure AI-generated content material. Related analysis means that AI disclosure might matter; throughout 13 experiments, Schilke and Reimann discovered that individuals trusted individuals and organizations much less when their use of AI was disclosed.

Evaluate AI Images Carefully Before Using Them

Organizations contemplating AI-generated imagery ought to consider every picture fastidiously reasonably than assuming that the standard of AI-generated photos and the way they are going to be perceived by customers might be inherently good or unhealthy.

Before utilizing an AI-generated picture:

• Check the aim: Does the picture help the web page’s message?
• Review illustration: Who is proven, who’s lacking, and the way are individuals portrayed?
• Assess authenticity: Do the individuals, setting, and interactions really feel plausible?
• Inspect for AI errors: Check textual content, palms, screens, glass reflections, and background particulars.
• Review it in context: Evaluate the ultimate crop, dimension, and placement within the interface.

The output itself can largely be evaluated utilizing the identical design standards as another picture. However, the methodology used to provide AI-generated imagery introduces extra issues.

Since image-generation fashions could also be skilled on images and different imagery depicting actual individuals, utilizing AI-generated photos raises questions on consent, resemblance, and the provenance of coaching knowledge. Teams ought to subsequently think about not solely whether or not an AI-generated picture works properly within the interface, but in addition the way it was created and the insurance policies and practices of the AI device used to provide it. Specifically, earlier than adopting an image-generation device, organizations ought to perceive its insurance policies round coaching knowledge and business use. An picture may fit properly from a UX perspective however nonetheless be inappropriate to make use of if its creation introduces copyright, consent, or different authorized and moral issues.

More About the Study Methods

The three AI-generated hero photos had been created utilizing ChatGPT Images 2.0 with the immediate: “Please generate 5 images that can be used as a hero image on a consulting company’s website where it shows people in action.” The inventory photos had been chosen from Unsplash and iStock utilizing a wide range of search phrases associated to office collaboration, akin to “business meeting collaboration” and “professional team meeting.”

To make the photographs fairly comparable, we utilized the identical common choice standards throughout each AI-generated and inventory photos. Each picture wanted to depict at the least three individuals collaborating in a office setting, present individuals in skilled apparel, be excessive sufficient high quality to be used as a web site hero picture, and have a broadly related colour palette in order that no picture stood out primarily due to its visible therapy. For the AI-generated photos, we additionally screened for apparent technology errors, akin to in human anatomy.

Participants considered 7 webpages in complete. One served as a follow trial and was all the time proven first; it was excluded from the evaluation. The remaining 6 webpages had been introduced in a randomized order.

Conclusion

Within the context that we examined, our findings counsel that the usage of AI-generated photos doesn’t essentially make individuals understand a web site as inherently much less reliable, skilled, or genuine. What mattered extra was the picture itself: whether or not it conveyed plausible interactions, acceptable illustration, collaboration, and a degree of visible high quality that match the group.

That doesn’t imply organizations ought to use AI-generated imagery with out scrutiny. AI photos nonetheless require cautious evaluate for inclusive illustration, realism, visible errors, and match with the encompassing content material. And as disclosure necessities turn out to be extra frequent, organizations must also think about whether or not realizing a picture is AI-generated might change how customers understand it.

Ultimately, the supply of a picture issues lower than whether or not the picture works for its supposed context — at the least when customers have no idea that the supply is AI. As with inventory images, deal with AI-generated imagery as one other design asset that should be evaluated fastidiously.

References

Le Monde with AFP. 2026. EU tells corporations to label AI-generated content material from Sunday. Le Monde. (July 28, 2026). Retrieved August 10, 2026 from

Oliver Schilke and Martin Reimann. 2025. The transparency dilemma: How AI disclosure erodes belief. Organizational Behavior and Human Decision Processes 188 (May 2025), 104405.