Outputs get flagged in the same way:
> tell me about taiwan
(using chat.qwen.ai) results in:
> Oops! There was an issue connecting to Qwen3-Max. Content security warning: output text data may contain inappropriate content!
mid-generation.