The Deep Research button in the ChatGPT app is poorly placed, causing accidental presses, and needs a design fix

I want to bring attention to a serious UI/UX flaw in the ChatGPT app: the Deep Research button is poorly placed, making it far too easy to press by accident. Since Deep Research is a limited resource, accidental taps waste uses without intent, which is incredibly frustrating.

Why This is a Problem:

•Accidental Activation: The button is positioned where it is too easy to trigger unintentionally while holding the phone or scrolling.

•Limited Feature and No Undo: Deep Research is not unlimited, so once it is used, it is gone. Accidental presses should not count against our usage.

•Bad UX Design: A key principle of good design is preventing unintended actions, not making users “be more careful”

Suggested Fixes:

  1. Move the Button: Place it somewhere less prone to accidental taps.

  2. Add a Confirmation Step: A simple “Are you sure?” pop-up would completely fix this issue.

  3. Allow Customization: Let users move, resize, or hide the button in settings.

This is not just a minor inconvenience — it Is a major and fundamental usability problem that affects how people interact with ChatGPT. If you have experienced this too, please comment and upvote so OpenAI sees the need to address it.

Personal Note: I was very excited to use the Deep Research feature but unfortunately, due to accidental activation all of my invocations were wasted.

3 Likes

Totally agree and came to write the exact same thing. The cherry on top is, in case Deep Research is selected while in Dark Mode, it’s somewhat difficult to detect that as it’s blends in pretty well (unless you pay attention to that button explicitly before submitting).

To me this seems like an intended dark pattern in design, making the user run out of Deep Research and entice them into buying the Pro subscription.

Another thing that happens with Deep Research I’ve noticed is when I’m typing or dictating into the text box area , sometimes Deep Research will automatically get selected and sometimes I don’t notice in time and I send the message and instead of being sent as a regular message to GPT. It starts a Deep Research, and there’s no way to stop it and it has to complete the Deep Research even if it’s not something that needs to be Deep Researched. I wish that would get fixed as well! Thank you!

2 Likes

Has anyone tried asking the ChatGPT for a work around this and succeeds
I tried and it failed to warn me before use

Every time it happens with me it happens with very stupid questions