The information displayed in the AIM should not be reported as representing the official views of the OECD or of its member countries.
A malfunction in AI-powered bot networks led to identical, AI-generated comments appearing under posts by Italian politicians, including Matteo Salvini, Carlo Fidanza, and Dario Nardella. The incident exposed coordinated use of AI bots to manipulate political discourse, raising concerns about interference and harm to democratic debate in Italy.[AI generated]
Why's our monitor labelling this an incident or hazard?
The event involves AI systems explicitly: generative language models connected via APIs to bot networks managing fake social media accounts. The malfunction (upgrade of AI models with stricter safety guardrails) caused the bots to post refusal messages instead of intended propaganda, revealing the coordinated manipulation. This use of AI bots to manipulate political discourse harms communities by undermining democratic processes and violates rights to fair political participation. The harm is realized and directly linked to the AI system's malfunction and use. Hence, it meets the criteria for an AI Incident rather than a hazard or complementary information.[AI generated]