University of Twente Student Theses
To Label or Not to Label : The Impact of AI Disclosure on Credibility and Engagement on LinkedIn
Schöning, Nils (2025) To Label or Not to Label : The Impact of AI Disclosure on Credibility and Engagement on LinkedIn.
|
PDF
3MB |
| Abstract: | AI-generated content (AIGC) is becoming increasingly common across social media platforms, and AI disclosure labels are emerging as a key mechanism in transparency initiatives. Yet, user interactions with AI labels remain largely underexplored. Therefore, this study investigates how LinkedIn users respond to content labelled as AI-generated, examining effects on sender credibility and content engagement. Moreover, it accounts for differences in content modality (text vs. image), and sender type (individual vs. organization), as well as the underlying factors social presence and user attitudes towards AI in content creation. Through a 2×2×2 experimental design, N = 237 participants evaluated fictional LinkedIn posts varying in AI disclosure and sender identity. The sample had a mean age of M = 31.1 years (SD = 10) and consisted of 58.2% female, 40.9% male, 0.4% diverse, and 0.4% non-binary participants. The findings revealed that AI labels impact credibility and engagement on LinkedIn in nuanced ways. While AI text labels reduce both sender credibility and engagement, labelling images show no significant effects. Meanwhile, social presence plays a central mediating role in shaping responses to AI-labelled content, whereas user attitudes towards AI moderate label-induced effects. Overall, the study reveals a tension between the increasing popularity of AIGC and user expectations of LinkedIn content. While labelling visuals as AI-generated does not cause meaningful effects, disclosing the textual component of a LinkedIn post as AI-generated jeopardises credibility and engagement. Ultimately, these findings underscore the importance of authenticity and originality on LinkedIn and highlight the need for platforms and users to carefully consider how AI is being integrated into the present and future content landscape. |
| Item Type: | Essay (Master) |
| Faculty: | BMS: Behavioural, Management and Social Sciences |
| Subject: | 05 communication studies |
| Programme: | Communication Science MSc (60713) |
| Link to this item: | https://purl.utwente.nl/essays/106122 |
| Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Show download statistics for this publication
Show download statistics for this publicationRepository Staff Only: item control page
