Echoes of influence: media systems and the representation of interest groups in artificial intelligence policy debates

Abstract

How do media systems shape which interests are heard in AI policy debates? We analyze 37,954 articles published between 2018 and 2024 in eight leading newspapers from Italy, Spain, the United Kingdom and the United States, using a large language model to identify the organizations cited and to classify the tone of their coverage. Three patterns stand out. First, economic actors dominate everywhere, appearing in roughly 85 per cent of articles, but liberal systems give noticeably more space to NGOs, academics and trade unions than polarized-pluralist ones. Second, the cross-national diversity gap does not come from how individual articles are composed; it comes from how often newsrooms publish multi-source stories at all. Third, sourcing profiles diverge across outlets in liberal systems and converge in polarized-pluralist ones, suggesting that market competition matters more than ideological alignment for editorial differentiation. Trade unions, notably, slip into negative territory in Spain.