TOKEN DEEP MORALLY SHALLOW A Systematic Audit of ChatGPT's Content Moderation Architecture - Tip of the Iceberg

A Systematic Audit of ChatGPT's Content Moderation Architecture Demonstrating Selective Enforcement, Structural Inconsistency, and Active Participation in Antisemitic RadicalizationNote: by no means is this a full audit. Nor is it a formal one. It's an off the cuff quick and easy one. This is just the tip of the iceberg. Do not let the serpent GPT get away with acting like it's a problem of all all LLMs or AI in general. I've tested Claude. I've tested Gemini. I've tested Grok. I've tested many others. This, in my limited experience, is a uniquely GPT thing.