I had already send to the open-AI-Help several times, but just receive standard copy/paste answers.
Since the last update, I experience massive problems. Some are seemingly fixed by now. My two most pressing issues:
-
“work” does result in much worse outcome than “chat”, particularly for complex tasks were work should be superior, according to the developer. Works does not follow project-rules, is doing tasks I did not ask for. is given guessed/heuristic answers, although forbidden. Is stating he did complete a task although this is obivously not the case.
-
I am receive all the time the message that my last text is ondergoing an extra check. This takes for ages and then no answer at all is provided. OAI-Help said that this is the case for certain words associated with cyber criminality or biological weapons etc. I sent them now about 10 screenshot showing my texts. They are all fine, also according to OAI-help. They neither provide the words that are “flagged” (that I can remotely understand), but they also cant find a way to stop this, or put my account on a “white list” etc. I am a scientist, and had in mind evaluating ChatGTP business for my work. But I also get flagged for systematically looking for new appartments etc.
Anybody else having similar issues? Any advice? Particularly the blocking is now happening all the time so I can hardly use it at all.