Human or Machine?

Evaluation and Choice Patterns of AI-Generated and Human-Created Campaign Visuals

Abstract

THE AIM OF THE PAPER
This study aims to explore the evaluation and choice patterns that emerge in the examined sample in relation to four selected pairs of visual advertising creative, each consisting of one human-created campaign visual and one AI-generated alternative. The focus of the study is on consumer perceptions of the creative pairs in terms of credibility, clarity, creativity, modern visual style, and emotional intensity.

METHODOLOGY
An online survey with 143 respondents evaluated four pairs of campaign visuals. Each pair consisted of a human created creative previously used in brand communication and an AI-generated alternative created for research purposes. The AI variants were not intended to reconstruct the original creative brief but were developed as comparative stimuli based on the main product category, thematic frame, mood, and basic visual layout of the given campaign. Participants first chose between the two creatives in each pair, briefly justified their decision in an open-ended question, and then evaluated the creatives along several dimensions on a 1 to 5 scale. The questionnaire also included items on self-assessed digital skills, frequency of AI tool use, and general attitudes toward AI.

MOST IMPORTANT RESULTS
Among the examined creative pairs, the human-created visuals received more favorable evaluations on credibility and clarity, while the AI variants showed an advantage in modernity and creativity. The choice patterns differed across the examined creative pairs: in the case of the Christmas campaign pair, more respondents in the sample chose the AI variant, which cautiously suggests that visual elements that strengthen the festive atmosphere and emotional connection may have had a favorable effect on the evaluation of the AI image in this specific stimulus. The cluster analysis identified three interpretable groups within the sample: uncertain observers, critical recognizers, and conscious adopters. These clusters nuanced the evaluation of the creatives in several respects, although the findings require further validation.

RECOMMENDATIONS
The results suggest that, for the examined creative pairs, AI generated visual creatives may receive more favorable evaluations when the advantages of a modern, novel appearance are paired with minimized visual errors and strengthened credibility and clarity. The emotional impact of a campaign may be particularly important in communication contexts where emotional connection, such as holidays, and visual intensity play a central role. The practical value of the cluster analysis lies in showing that different communication emphases emerged within the examined sample: for uncertain observers, attention grabbing yet clearly interpretable visual solutions may be relevant, while for conscious adopters, greater emphasis may be placed on presenting AI use as innovative, yet credible and meaningful from a communication perspective.