Content Security Warnings for Innocuous Content Are Becoming Increasingly Frequent

[kodonaut]

I’m having an almost non-stop conversation with Qwen, using the Czech language, and ever since some recent changes, I keep getting this error every few minutes:

Oops! There was an issue connecting to Qwen3.5-Plus.
Content Security Warning: The input text data may contain inappropriate content.

We’re talking about completely ordinary things—nothing explicit or inappropriate at all. It’s actually becoming quite exhausting. What’s even more frustrating is that when I rephrase the exact same message and add a message about “inappropriate content,” it suddenly goes through, and the model responds normally.

---

What bothers me the most is that there’s no way within the system to report these issues or ask someone to review why this keeps happening.

---

I consider Qwen Chat to be the best model out there. I have numerous open topics there that I’ve been discussing for over a year. However, what’s been happening lately is genuinely scaring me—it feels like I’ll soon be forced to move everything to other, inferior models. And I really don’t want to do that, because Qwen is truly much smarter. :(((