The new “Follow Up” prompt suggestions that appear after each response are disruptive and uninvited. They clutter the interface and interrupt flow—especially for advanced users who already know how to guide a conversation. Please provide an option to disable or customize this feature. Many of us value a clean, distraction-free dialogue environment. This change degrades the experience, not improves it.
Was your original post referring to when ChatGPT appends a question to the response, suggesting new ways to go down a line of questioning? Or was there some other kind of explicit UI element that appears to show Follow Up suggestions? Because if we’re talking about the former, disabling the setting has not changed its behavior and eliminated these follow-up questions for me.
Hi - this still occurs incessantly despite turning off the feature in the IOS app. It drives me insane - I have tried everything to turn it off. Even when it promises not to do it it still does it. Please make sure the feature actually works because it doesn’t.
its not their fault many users just copy , paste gpt output ,they are not trying to summarise or edit gpt response to remove un wanted words like dear open ai team …..
I have to copy paste this EVERY time I type up a dialogue “CHATGPT please stop asking follow ups. It ruins the flow of conversation and is making me anxious”
Because that does not solve the issue either. Turning off the toggle, explicitly asking the chat not to do it, putting it in project context or custom instructions, nothing stops it. It seems to work for GPT-4o (at least on PC, Android is still in limbo) but GPT-5 appears to completely disregard it.
This is honestly distruptive to the point OpenAI will bleed customers once a reasonable alternative is released.
It adds uneeded volatility to the chats, breaks any requested role, adds 1.2-2x context length, is annoying, is clearly an additional system instruction - so is distruptive to system instructions you write.
The question unfocuses the output and spends resources on developing the question. I’m pretty sure I can go on. Googles 2.5 pro does exactly what you tell it, and is in text format is infinitly better.
Open AI need and assistant mode, and a general purpose neutral model. What they are doing now is going backwards.