Advanced Voice upgrade ruined the flow of useful voice chats for me on Android, can I downgrade?

While I do like the new voices, my primary use case for the voice mode in ChatGPT on my phone (Android) was to discuss pre-uploaded papers, and search the web etc. while driving to work or while doing other chores that required my hands and visual attention.
(I do not care about cute voices, vocal emotional fakery and other voice gimmicks).

The problem is that after the “Advanced Voice” upgrade, the “tap to interrupt” function does not work on my phone any more - anytime I tap the screen to interrupt the speaking of the chatbot, so I can ask a new question, the conversation cannot continue, because ChatGPT becomes unresponsive, as if it cannot “hear” me anymore, and nothing happens after that.

The app is not frozen overall, I can still close the chat and start it again, but it’s as if the app cannot receive any input from me (even though there is still some “listening” animation on the screen, it does not seem to react to my voice and questions).

Checking the chat transcript after that reveals no events after I tapped the screen - it’s as if the app muted me when I tapped the screen, so I could not ask anything anymore. The only solution is to tap the circled x button to close the conversation, then tap the even smaller voice chat icon in order to restart the conversation (which sometimes takes a while to initialize, for whatever reason…).

This is breaking the pseudo-hands-free voice chat flow I had before, and it’s very inconvenient and potentially dangerous to do all that extra tapping while driving (tapping the small voice icon again, specifically).

Ideally we should be able to have a fully hands free experience, in Standard Voice mode - how hard it is to interrupt the chatbot with our voice (with a loud stop word or something), why is this not implemented yet (for really useful conversations, with file uploads and web access) ?

(Gemini live CAN chat with documents fully hands free, but as we all know it’s otherwise dumber and a bit obtuse, so not really useful, and it cannot search/access the web… despite being made by Google… go figure…).

4 Likes

I have a similar problem. I was using a custom gpt i created to chat about uploaded docs in the knowledge. I now get a message saying my current chat has content that advanced voice mode " cant handle just yet". I can’t seem to downgrade to the non advanced version which renders my gpt useless. Any ideas?

1 Like

I guess I get a different message at the bottom as I enter voice mode in a pre-prompted conversation, which I am happy to ignore since I do not need those fancy voice features…:

My understanding was that this is the new “standard” voice mode for such pre-prompted conversations (or in custom GPTs as you observed, though you do seem to get a different message), and non-advanced would be perfectly OK for my use case IF the tapping wouldn’t completely “mute” me to the app from then on!

By the way, before I tap on the screen , the voice is of impressive quality (using Vale), I like it so much more than the voices we had before, the inflexions sound very “advanced” to me…
If it only had the interrupt feature working (preferably by voice, it cannot be that hard ! but I’ll take the tapping too…), then it would be a really fantastic experience.


If you click on the little “i” on the message you screenshot you get the message i referred to. Seen in the attachment. Im going to try creating a new gpt with the same criteria and knowledge files and see if that helps.