The Media Career of Artificial Intelligence in Europe’s Elite Press, 2015–2026

Abstract

Artificial intelligence (AI) has become a fixture of the news agenda, yet little comparative research traces its coverage across countries and across the generative turn marked by ChatGPT's release in late 2022. This study analyzes 43,986 articles from leading center-left and center-right newspapers in ten European countries and seven languages from 2015 to mid-2026. The articles were coded by an ensemble of large language models and the resulting classifications were validated against human judgment. Coverage grew, first peaking in the late 2010s and then surging far more dramatically after the release of ChatGPT, reaching more than seven times its earlier average level by the first half of 2026. AI coverage became markedly more negative, as net valence fell from +12.5 in 2022 to -21.7 in the first half of 2026. The decline occurred in all ten countries and among both center-left and center-right newspapers, without a consistent ideological pattern across countries. What made AI newsworthy also changed. Science and research fell from 26 to 12 per cent of coverage, while stories about disputes over copyright, jobs, and chatbot harms became more prominent. In its media career, AI thus shifted from an emerging technology covered largely through capabilities and potential applications toward a technology in use, increasingly covered through its consequences, conflicts, and governance.