Facebook's advertising algorithms continued to serve ads for military gear and weapons accessories alongside extremist content and election misinformation in militia groups after the January 6 Capitol riot, despite the platform's stated policies against such advertising.
Following the January 6, 2021 Capitol riot, Facebook continued to display advertisements for military gear, weapons accessories, and body armor alongside extremist content and election misinformation in far-right militia groups. The Tech Transparency Project (TTP) documented this through a monitoring account that followed over 75 pages or groups sharing pro-Trump content, election misinformation, and calls for violence. Facebook's algorithms served ads for gun holsters, rifle attachments, and tactical gear to users engaging with content from groups like 'RAPID Militia,' 'Modern Militia,' and 'Appalachian Mountain Patriots.' Internal Facebook employees flagged these ads as problematic, with one noting the 'terrible timing' of advertising military gear alongside rifle imagery. On January 15, 2021, senators and state attorneys general called on Facebook to halt such advertising. Facebook announced on January 16 that it would ban ads for weapon accessories and protective equipment through January 22, but TTP continued finding such ads even after this announcement. The incident highlights how Facebook's ad targeting algorithms profited from extremist content by serving militaristic advertisements to users consuming violent and conspiratorial material.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
Using AI systems to conduct large-scale disinformation campaigns, malicious surveillance, or targeted and sophisticated automated censorship and propaganda, with the aim of manipulating political processes, public opinion, and behavior.
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