Skip to content
LLMMedium

Twitter's AI Moderation Tool Misidentified Rockets as Pornography

Occurred
Jan 3, 2023
Published
Jan 5, 2023
Source
AI Incident Database

Summary

Twitter's automated content moderation misidentified images of rocket launches as pornographic content, prompting incorrect account suspensions.

Developers

  • Twitter

Deployers

  • Twitter

Harmed parties

  • X (Twitter) users

Classification

Category confidence: 85%
Severity confidence: 90%

Rationale: Automated moderation system error causing account suspensions due to misclassification.

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

Visit source