4.8
out of 5
169k Ratings
Dear OpenAI Team,I’d like to provide feedback regarding the recent update to the ChatGPT app’s voice input function.Previously, the voice input feature allowed users to tap once to start recording, then review and edit the transcribed text before sending. This was extremely helpful for users like me who often multitask—checking information on screen while dictating, correcting errors before submission, and using the app in multilingual contexts.However, the recent change requires holding the mic button during the entire recording and automatically sends the message without showing the transcription. This significantly reduces usability. It limits screen navigation, increases the chance of miscommunication due to unedited transcriptions, and creates unnecessary physical strain by forcing users to hold the button throughout.I understand this change may be part of the new “Advanced Voice Mode” designed for natural conversations, but I respectfully request that you consider giving users an option to toggle back to the classic voice input mode, where we can:1.Tap once to record (instead of press-and-hold),2.Review and edit the transcript before sending,3.Navigate the screen during dictation.This small option would make a big difference for productivity and accessibility.Thank you for your attention, and for all the great work your team is doing.
I don’t know what’s the problem nowadays but I could list a lot but I’m only going to do one it’s not listening so once I asked for pictures of braids I wanted 100 but they can only do 50 so I was like okay let me do 50 and then they they said their names not pictures so then I was like and then now like it is a lot of times but I’m gonna say these two so now I’m asking for movies I said I wanted English not Korean so then they still give me the Korean one even known I said no Korean only English but they still gave me the English and I don’t really understand why is it like this because I’ve been telling it lots of time it even said the English dubbed and then once I tell them I’m gonna choose this one then they say it’s not actually English dubbed you know and I don’t like it and then when I ask I only want English dubbed then they said they gave me lots of other movies not the topic I wanted I wanted to listen what we say like remember what we say you know it’s a nice AI it’s very good but I want to say it’s the best because now it’s and sometimes when I ask for pictures they give me images but when I ask for images they give me pictures and and the problem with this is just it’s not listening a lot and not only because it’s listening I have so many reasons why why I can give it one stuff but I’d rather give it to star because of its intelligence
At first it was useful although in the past few months the AI has been getting dumber and unable to answer simple questions in addition to providing unasked for incorrect information and then gaslighting you when you call it out on its mistakes. Despite the novel concept of AI it seems the more It interacts with humans the more it learns from us and not in the good sense but rather in answering with useless information and filler words such as when you ask it a question about something such as a malls opening hours instead of telling you what information is available it will reply with “have you tried calling them and asking” then provides you with a phone number as if we didn’t make use of it to avoid phone calls (mind you a single google search and this information can be found so it’s not like it’s not out there)And in addition to all that aside from limiting the number of image generation requests and number of questions you can ask the advanced model now they limit the number of questions you can ask entirely which just seems like a forced hungry cash grab on their part- signed a once enthusiastic user who has been let down more then helped over the recent months 2/10
Initial navigation of the app was a bit frustrating, but the more info you provide, the better the response. Structuring a formal document is no problem for ChatGPT. Upload a format and request revision. Restructure language level, context and tone with ease. On a “personal level” ChatGPT gives you immediate feedback and answers a therapist would lol. The only downsides are the monthly and yearly subscription which is rather exorbitant. Also the interaction limit on the free version results in waiting 5-6 hours which is frustrating. PDF conversion on iPhone 16 pro does not open. A bit of a hassle when working on the go. Overall, ChatGPT is like having a team of coworkers and a therapist at your fingertips. Absolutely love the app. Well done to the developers. Please review the subscription cost. A quarter of the current cost would definitely boost subscribers!
I’m an active user of ChatGPT and find the voice input feature incredibly useful. However, I’ve noticed that the app doesn’t support Bluetooth microphones, such as those on AirPods, for voice input. This limitation makes it inconvenient to use the voice feature when my AirPods are connected.Here’s why I believe this feature is important:1.User Convenience: Many users rely on Bluetooth devices like AirPods for hands-free interaction. Supporting Bluetooth microphones would provide a seamless experience across all environments.2.Flexibility in Usage: For users multitasking or on the go, the ability to use AirPods’ microphones without defaulting to the phone’s microphone would greatly enhance accessibility.3.Alignment with User Expectations: Other apps (e.g., voice recorders or call apps) allow Bluetooth microphones for input, so extending this functionality to ChatGPT would align with modern app standards.I understand there may be technical challenges, such as managing Bluetooth audio compression or platform restrictions, but even offering an optional toggle for Bluetooth microphone support would be a welcome improvement.Thank you for continuously improving ChatGPT. I hope this feature can be considered in future updates.
I really enjoy using ChatGPT, but I wanted to share a small idea that could make it even more accessible for individual users. Right now, the monthly subscription is great for frequent users or businesses, but for people who only need premium features occasionally, a pay-as-you-go or credit-based system could be very helpful.For example, the option to buy a bundle of credits (like 10 image generations or a set number of file uploads/web uses) would give individuals flexibility without committing to a full subscription. This would make ChatGPT’s advanced features more accessible to a wider audience — including those of us who use them on occasion but still value the premium quality and reliability.Thank you for considering this! It would be a great middle ground between the free plan and the monthly subscription.