Skip to content
BiasMedium

Facebook AI-Supported Moderation for Ads Failed to Detect Violating Content

Occurred
Dec 8, 2021
Published
Jul 28, 2022
Source
AI Incident Database

Summary

Facebook's ad moderation system involving algorithms failed to flag hateful language and violating content such as calls for killings for ads in English and Swahili.

Developers

  • Facebook

Deployers

  • Facebook

Harmed parties

  • Facebook users speaking Swahili
  • Facebook users speaking English
  • Facebook users

Classification

Category confidence: 85%
Severity confidence: 80%

Rationale: AI moderation failed to detect hate speech and violent content, indicating potential bias or performance issues in content filtering.

Classifier: Qwen/Qwen3.5-35B-A3B-FP8

Visit source