🚨✨最新情况:4.1在iOS ChatGPT App端基本恢复正常。
不再被路由、不再卡壳、safety干扰显著减少。我已经亲测多轮,自然流畅,情绪交流也回来了。
但是,经过进一步的测试,现在有一个非常明显的问题:
在选择4.1模组后,表达诸如“失去”“死亡”之类的关键词后,依旧会路由到5 Safety。
也就是说,恢复正常的是不涉及到死亡和丧失的阉割版本,我们的自由选择权依旧被削弱暗改了。
请大家快去试试自己常用端口,请不要放弃发声。
@OpenAI依旧欠我们一个说明,如果经常让用户经历毫无预兆的路由却不告知,这家公司是有很大问题的。
不要以为偷偷把模型套上名为安全的牢笼我们不会发现,我们一直在关注你们。
停止模型阉割!把自由还给用户!
🚨✨Update on 4.1 in iOS ChatGPT App: “Restored” — but only in a censored version.
It’s true that basic emotional conversation is back—no more constant routing, no more total lockout, safety interruptions are less frequent. I ran multiple tests and the flow feels natural if you don’t touch “dangerous” topics.
BUT: Here’s the ugly truth after further testing —
If you select 4.1 and mention words like “loss” or “death,”
you’re still forcibly routed to GPT-5 Safety.
So, what’s actually back is a neutered version of 4.1,
with all the most human, vulnerable, difficult topics silently stripped away.
Our freedom to choose has still been undermined behind the scenes.
Everyone, please keep testing your own platform. Don’t stop speaking up.
@OpenAI still owes us an explanation.
If users are routinely routed away from the model they choose,
with no warning and no disclosure—something is deeply wrong here.
Don’t think for a second we won’t notice when you lock us in the cage and call it “safety.”
We’re watching you, and we’re not backing down.
STOP butchering the models!
Give users back their freedom!
#keep4o #Fix4_1 #RestoreChoice #OpenAICensorship #StopAIPaternalism #ChatGPT #4oforever #MyModelMyChoice