FAIL · AI INCIDENT TRACKER

wire / 2026-08-19-d5ad

Meta ad moderation kept running deepfake scam ads impersonating Polish businessman Rafal Brzoska

August 19, 2026 · POL INCIDENT ECONOMICRIGHTS

Meta kept serving the ads after a court told it to stop. The court order appears to have been a suggestion.

The facts

Meta's AI-driven content moderation failed to block deepfake scam ads on Facebook and Instagram that illegally used the images of Polish businessman Rafał Brzoska and his wife, leading to financial fraud and privacy violations. Despite court orders and regulatory demands, the harmful ads persisted, highlighting weaknesses in Meta's AI systems.