The model’s flawed responses to prompts involving Tibet, Taiwan, and Falun Gong raise red flags about the influence of state-aligned censorship on AI reliability, with experts calling for stronger ...