RISK SOCIETY IN REVERSE: NEWS COVERAGE OF ARTIFICIAL INTELLIGENCE, 2019-2025

Abstract

The recent rapid spread of artificial intelligence (AI) applications, including facial recognition-powered surveillance, chatbots, and AI agents, has triggered broad economic and social changes. Many of these are harmful, including job losses and economic dislocation; the capacity to produce mis- and disinformation at scale; impacts on cognition, learning, and mental health; algorithmic bias against women and people of color; expanding surveillance nets and loss of privacy; and data security risks. Members of Silicon Valley’s elite say that AI will transform society for the better, but some of them also warn that a rogue superintelligent machine could wipe out humanity. Unlike many other technologies, AI is mostly unregulated in the United States, leaving it an open question of how society should respond to these hazards. Journalism occupies a central role in addressing that question: its mission includes informing the public about technological risks and ensuring accountability from powerful actors. At this moment of potentially radical technological change, how is it performing that key role of mediating public understanding of risks from AI? To explore that question, this study employs a qualitative content analysis of critical discourse moments to analyze how eight news organizations – the Associated Press, New York Times, Wall Street Journal, Washington Post, TechCrunch, TechRadar, VentureBeat, and Wired – covered AI from 2019 to 2025. This was a period of political and social upheaval in which forms of AI, including facial recognition technology and large language models (LLMs), became increasingly intertwined with news events. These included the George Floyd protests in 2020, the emergence of an alliance between tech CEOs and President Donald Trump, and efforts in 2025 by the Elon Musk-led Department of Government Efficiency (DOGE) to disrupt federal agencies. Employing Ulrich Beck’s risk society framework and Pierre Bourdieu’s field theory to map how news stories cover technology risks and the social, political, and economic forces acting on news organizations, the study finds that journalists at times exercised their accountability role effectively, exposing AI hazards to public scrutiny and stimulating debate about them. But they failed during the initial rollout of generative AI in 2023 and generally struggled to cover AI risks consistently amid intense pressures, including industry hype, ongoing stresses on the news business, attacks on the media by Trump and the MAGA movement, and the seeming unraveling of U.S. risk management institutions themselves. The findings suggest journalism’s essential role in exposing technological hazards has weakened, casting doubt on whether basic democratic processes are robust enough to contend with the risks of expanding AI systems. By applying this Beck-Bourdieu-critical discourse moment framework to news coverage of rapidly evolving AI debates, this study enables insights into news organization framing of AI during a crucial period of tech industry ascendancy and U.S. political disruption; examines contradictory news priorities of promoting technology and progress vs. public accountability; and shows how the tech industry’s increasing power, news industry weaknesses, and erosion of risk norms manifest in news content and public debates over AI. This study also makes recommendations for journalism practice, specifically on how journalists can more effectively cover risks from AI and other technological systems, contributing to stronger public accountability.