Abstract
This bachelor’s thesis examines the perception of AI-generated advertising images on social media. The relevance of this topic arises from technological advancements in the field of generative artificial intelligence. In particular, text-based image generators enable companies to create visual advertising content more quickly and cost-effectively. At the same time, new challenges arise, as AI-generated images may be viewed more critically by consumers, thereby potentially affecting credibility, trust, and authenticity. Additionally, Article 50 of Chapter 4 of the EU AI Act will lead to consumers being explicitly informed of the artificial origin of such content in the future.For the theoretical evaluation, the fundamentals of generative artificial intelligence and its potential applications in marketing were first presented. Social media was then described as an advertising environment characterized by various attention challenges. Furthermore, various theoretical approaches to advertising perception as well as specific characteristics of AI-generated advertising images were examined. This included a focus on AI hyperrealism, the Uncanny Valley of Mind, the lack of emotionality, and the transparency label. The empirical study was conducted using a quantitative, exploratory-descriptive research design. Using a standardized online questionnaire, participants evaluated three AI-generated advertising images labelled as “AI Generated.” The advertising products used were a notepad, toothpaste, and a blanket. The final sample comprised N=102 valid cases. The evaluation was based on the three constructs of credibility, trust, and authenticity.
The results show that AI-generated advertising images were perceived by survey participants as generally neutral to slightly positive. The overall mean score for credibility was M = 3.25, for trust M = 3.24, and for authenticity M = 3.11. The standard deviations ranged between 0.91 and 1.16. It was found that authenticity was perceived most critically compared to credibility and trust. A comparison of the three stimulus images also illustrates that it was not the AI label, but rather the specific image design that had a significant influence on perception. The emotionally staged blanket image achieved the highest scores across all dimensions, while the toothpaste image with the artificially generated human face received lower ratings, particularly in terms of authenticity. The results suggest that AI-generated advertising images can generally be used in social media marketing, but their success depends on a natural, credible design that aligns with the product, as well as a transparent labelling.
| Date of Award | 2026 |
|---|---|
| Original language | German (Austria) |
| Awarding Institution |
|
| Supervisor | Stephanie Bauer-Pramhas (Supervisor) |
Studyprogram
- Marketing and Digital Business
Cite this
- Standard