I’m writing to express my strong dissatisfaction with the recent changes to the voice-to-text feature in the ChatGPT app. The automatic sending of dictated text without the ability to revise and edit is a significant step backward in usability and functionality.
Previously, the ability to review and correct transcribed text before sending was crucial for:
Accuracy: Voice recognition is not perfect, and errors are common.
Clarity: Editing allows for the refinement of phrasing and grammar.
Privacy: Reviewing text before sending ensures sensitive information is not accidentally transmitted.
Control: Users should have the final say in what is sent.
The current implementation forces users to either accept potentially inaccurate and poorly phrased input or abandon voice dictation altogether. This severely limits the usefulness of the feature, especially for users with disabilities or those who prefer voice input.
I urge the OpenAI team to:
Revert to the previous functionality, allowing users to edit voice input before sending.
Alternatively, provide an option to enable or disable automatic sending. This would allow users to choose the workflow that best suits their needs.
This change has negatively impacted my experience, and I believe it has done so for many others. I hope you will take this feedback seriously and prioritize user control and accuracy.
Thank you for your attention to this matter.
I just opened another thread without knowing this one existed. I also consider this to be a serious problem and a significant regression in terms of usability. I hope they read this thread or the one I created, because I think it’s an important issue.
Agreed wholeheartedly. I just want to chime in and say I strongly prefer the old version of voice-to-text. This new setup is a major downgrade. It feels less intuitive, more intrusive, and honestly, it disrupts the creative process. The previous version let me speak my thoughts, see them transcribed, tweak things if needed, and then send…that workflow made sense.
Now, it just sends everything the moment I stop talking, which is clunky and makes me feel rushed. It removes that crucial review/edit step and splits up the flow in a way that kills spontaneity and natural expression. I really dislike this change, and I hope there’s a way to revert to or restore the older experience. This current version is basically unusable compared to how it was previously.
I find this new change extremely impractical and, in many situations, even counterproductive. The previous option to convert longer dictations into text in stages was a key advantage—especially for complex inputs or extended prompts.
Previously, after about two minutes of speaking, you could press the button to convert the spoken content into text. You then had the opportunity to edit it, add more content, or dictate another section before sending everything together. This wasn’t just convenient—it was also a crucial safeguard against data loss.
Now, however, the text is sent immediately once you tap the arrow button—without any chance to review or add to it. Step-by-step dictation is no longer possible. Even more concerning: should the speech go beyond three minutes, as sometimes happens to me, a network error occurs and the entire text is lost. There’s no backup, no recovery, and no warning.
This change significantly limits the practical use of voice input and makes longer, well-thought-out speech-to-text entries virtually impossible. I sincerely hope that OpenAI recognizes this issue and provides an option to restore the previous behavior.
Upvoting this request. This new feature is terrible. We need to be able to view the Voice to text transcript before sending it over. They should have given us the option to enable or not this new feature.
2 weeks later and many of us are still stuck with this core feature broken. A simple error in transcription can derail the entire conversation and requires backtracking. Being able to check the transcript first is absolutely essential.