AI-Generated Disinformation in the UK: Risk of Harm, Context, and Classification
This study addresses the threats posed by AI-generated disinformation to social trust and public safety by presenting the first systematic quantification of its multidimensional harm landscape within UK society. Through the analysis of 112 high-exposure instances of AI-generated false content, this work employs an empirical data-driven taxonomy and risk assessment methodology to identify substantive risks across eight major domains. The findings demonstrate that AI-generated disinformation significantly exacerbates the risks of social unrest, health-related harms, and financial fraud, while broadly eroding public trust in information ecosystems. By mapping these vulnerabilities, this research provides critical empirical evidence to inform effective AI governance frameworks and targeted mitigation strategies against the escalating societal impacts of synthetic media.