The information displayed in the AIM should not be reported as representing the official views of the OECD or of its member countries.
OpenAI reported to Brazil's Federal Police and electoral authorities that at least two ChatGPT users expressed violent threats and intentions to harm President Lula. The company shared chat logs and closed accounts, prompting potential criminal investigations. The AI system was directly used to articulate and document these threats.[AI generated]
Why's our monitor labelling this an incident or hazard?
The AI system (ChatGPT) was explicitly used by the individual to plan and discuss violent acts, including assassination and mass violence, which are direct threats to human life and safety (harm to persons). The misuse of the AI system to facilitate these threats constitutes a violation of legal frameworks and human rights protections. OpenAI's intervention and notification to authorities confirm the seriousness of the incident. Given the direct involvement of the AI system in enabling these harmful communications and the realized threat (even if the attack has not yet occurred, the threat itself is a harm), this qualifies as an AI Incident rather than a hazard or complementary information.[AI generated]