This publication investigates the construction of AI narratives within the European press, analysing how “success” and “failure” were framed between 2022 and 2025. Through a computational analysis of 31,294 articles across France, Germany, Ireland, and Spain, the research maps the themes, patterns, and temporal trends related to AI success and failure, identifying key topics, terminology, and narrative structures.
The analysis identifies eight cross-country thematic clusters that structure press discourse on AI and its evaluation: international and geopolitical developments, economy and finance, health and medical innovation, regulation and legal frameworks, culture and creative industries, training and education, cross-sectoral uses of AI, and cross-cutting risk-oriented concerns
The data also reveals a media discourse dominated by powerful industry and political actors, while civil society perspectives remain largely marginalised. The publication shows that news media tend to frame AI success primarily through economic and technological lenses, revealing a tension between competitiveness and expectations as regards ethics and fundamental rights, shaping how AI applications are understood and socially accepted in Europe.
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
