Made by a Machine, Liked by a Human? AI-Generated vs. Human-Created Content in Social Media Marketing and the Effect of Disclosing “Made with AI”
0 viewsAbstract
Scroll through Instagram today and there is a good chance that at least one of the posts you liked was not made by a person. Brands are turning to generative AI tools such as ChatGPT, Midjourney and Canva’s Magic Studio to produce captions and visuals faster and cheaper, but we still know very little about how consumers actually react to this content, especially when they are told it was made by AI. In this study I tested whether AI-generated social media posts perform differently from human-created ones, and whether a “Made with AI” label changes the picture. I ran a between-subjects experiment with 120 USEK students who were randomly shown one of three versions of the same Instagram campaign for a fictional Lebanese coffee brand: human-created, AI-generated without disclosure, and AI-generated with disclosure. I measured four outcomes: engagement intention, perceived authenticity, brand trust and purchase intention, and compared the groups using one-way ANOVA. The results show that when participants did not know the content was AI-made, they rated it almost the same as the human-made posts, and even slightly higher on engagement (3.66 vs. 3.58). Once the label was added, however, every score dropped, with perceived authenticity falling the most (from 3.81 to 2.74). In other words, the problem is not the AI content itself but what people think when they find out. These findings suggest that marketers can use AI to support content creation, but that transparency rules, such as the EU AI Act, will force brands to think carefully about how they communicate AI use to their audience.
Keywords
Ai Disclosure, Brand Trust, Consumer Behavior, Generative Ai, Perceived Authenticity, Social Media Marketing