When AI Models Police Your Social Media Speech

In episode 25 of AI Ethics with Fexingo, Lucas and Luna explore the hidden world of AI content moderation on social media platforms. They zoom in on a 2025 study from the Algorithmic Justice League that found automated moderation tools disproportionately flag speech from minority communities—misinterpreting dialect, slang, and cultural context as hate speech or misinformation. Lucas breaks down the specific numbers: how a model trained on mainstream English has a 40 percent higher false-positive rate for African American Vernacular English. Luna questions whether transparency reports are really making a difference when the underlying training data remains opaque. They discuss the tension between keeping platforms safe and preserving free expression, and what it means when an AI with no understanding of nuance decides what you can and cannot say online. The episode is grounded in real examples—including the 2024 controversy over TikTok's moderation of LGBTQ+ content in the Middle East—and closes with a look at emerging audit frameworks that could shift the balance toward fairness. #AIEthics #ContentModeration #AlgorithmicBias #SocialMedia #FreeSpeech #HateSpeechDetection #Misinformation #AfricanAmericanVernacularEnglish #AlgorithmicJusticeLeague #TikTok #LGBTQ #Transparency #TrainingData #AuditFramework #Technology #FexingoBusiness #BusinessPodcast #Fexingo Keep every episode free: buymeacoffee.com/fexingo