Model1h ago
AI Chatbots Exhibit Persistent Left-Leaning Bias
GPT-5.5 provided exclusively left-leaning arguments 80 percent of the time in a Washington Post investigation. Even Grok skewed left despite its anti-woke marketing. Google's Gemini 3.1 Pro stood out by presenting both sides 93 percent of the time. These results highlight the ongoing difficulty of achieving political neutrality in LLM training.