Instagram's algorithm malfunctioned and flooded users' Reels feeds with violent and graphic content including videos of people being killed, dismembered, and tortured, exposing users to disturbing material that violated the platform's content guidelines.
Meta's Instagram experienced an algorithm malfunction that caused users to receive violent and graphic content on their Reels feeds. Users reported seeing videos of people being killed, dismembered by helicopters, crushed by elephants, set on fire, and other extremely violent content. One user reported seeing 'at least 10 people die' on their reels, while another described their feed as 'absolutely packed with violent Reels - like, one after the other.' The malfunction affected users across the platform, with many reporting content from accounts like 'PeopleDeadDaily' appearing in their feeds despite having unrelated interests like biking. Meta acknowledged the error and apologized, stating they had 'fixed an error that caused some users to see content on their Instagram Reels feed that should not have been recommended.' The company clarified that this incident was not related to their recent changes to content moderation policies. The incident raised concerns from safety groups like the Molly Rose Foundation about the presence of such graphic content on the platform and its potential impact on users, particularly given Instagram's history with harmful content exposure.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
AI that exposes users to harmful, abusive, unsafe or inappropriate content. May involve providing advice or encouraging action. Examples of toxic content include hate speech, violence, extremism, illegal acts, or child sexual abuse material, as well as content that violates community norms such as profanity, inflammatory political speech, or pornography.
AI system
Due to a decision or action made by an AI system
Unintentional
Due to an unexpected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed