This publication examines how public discourse on social media negotiates, shapes, and contests the definitions of AI “success” and “failure.” Drawing on a comparative dataset of user-generated content from Facebook, TikTok, and YouTube across France, Germany, Ireland, and Spain, the research applies lexicometric and network analysis to identify recurring societal tensions.
The findings indicate that societal acceptance of AI is largely conditional and instrumental. Users frequently frame “success” through the lens of utility, adopting Generative AI tools to enhance professional productivity, scale content creation, or complete educational tasks. Conversely, “failure” is predominantly framed as social disruption, with discourse focusing on labour displacement, the proliferation of deepfakes, and the erosion of human agency. Many users adopt Generative AI to “keep up” with platform economies or workplace expectations rather than because they endorse its broader trajectory.
The analysis also reveals distinct platform dynamics. Facebook functions as a space for institutional and semi-professional discourse, TikTok fosters a utilitarian environment focused on algorithmic visibility and efficiency “hacks”, and YouTube serves as an arena for elaborated debates. By mapping these framings, the publication highlights the specific conditions under which European citizens grant or withhold legitimacy to AI technologies.
This publication can be downloaded from Zenodo. The publication hasn’t yet been reviewed and approved by the European Commission.
Authors: Charis Papaevangelou, Lucie Loubère, Nikos Smyrnaios, and Pierre Ratinaud
