
The information displayed in the AIM should not be reported as representing the official views of the OECD or of its member countries.
Facebook's AI-driven moderation systems mistakenly disabled thousands of user accounts globally without warning or clear explanation, denying access to personal data and memories. Meta acknowledged the issue and is investigating, but many users remain locked out, highlighting the risks of automated AI enforcement in social media platforms.[AI generated]
Why's our monitor labelling this an incident or hazard?
Facebook's automated content moderation and account enforcement systems use AI to detect violations and take action such as disabling accounts. The sudden, unexplained mass account lockouts indicate a malfunction or misuse of these AI systems, leading to harm in the form of loss of access and potential violation of user rights. The harm is indirect but significant, affecting many users and their ability to use the platform. This fits the definition of an AI Incident because the AI system's use has directly or indirectly led to harm (violation of user rights and community disruption).[AI generated]