Informing AI Risk Assessment with News Media: Analyzing National and Political Variation in the Coverage of AI Risks

AI risks are not universally prioritized: The study found that different countries emphasize different AI risks in their news coverage, suggesting that a one-size-fits-all approach to AI governance may miss culturally and nationally specific concerns that matter to different societies.

Political orientation shapes risk narratives: In the U.S., left-leaning and right-leaning news outlets not only prioritize different AI risks but also use "politicized language" when reporting on them, indicating that the same risks can be framed very differently depending on political perspective—which may influence which risks policymakers ultimately address.

News media reflects societal context, not just technology: The paper argues that focusing only on the technical features of AI systems misses "systemic risks that emerge from the complex interaction between AI systems and society," and news media can help capture these broader societal dimensions that pure technical risk assessments might overlook.

Media coverage influences which risks are considered important: Because news media has "agenda-setting" power in shaping public discourse, the variations in how different countries and political outlets cover AI risks directly influence which risks are deemed important enough for policy attention—making media analysis relevant for governance decisions.

Risk assessors should account for media bias when using news as evidence: Policymakers considering news media as a source for identifying AI risks should recognize that coverage varies by country and political orientation, requiring careful interpretation rather than treating media reports as objective reflections of actual risk severity.