ChatGPT dictation now shows text as you speak, and the 10-minute limit is gone
OpenAI is bringing ChatGPT a dictation feature, Live Dictation, that turns speech into text while the user is still speaking. The previous 10-minute recording limit is being removed, and the feature is rolling out in stages to all plans, including the free one.
In release notes published on 18 November, OpenAI said it is gradually rolling out a ChatGPT dictation feature that shows the transcribed text while the user is still speaking. South Korea’s AI Times reported the news. Previously, users had to wait until the whole recording had been converted into text. Tapping the waveform in the input field shows or hides the text.
Dictation is separate from voice conversation, in which ChatGPT replies by speaking. In dictation, the user taps the microphone icon and their speech becomes a written message, which they can check and correct before sending. AI Times said this lets someone dictating a long question or draft spot missing parts and recognition errors while still speaking.
OpenAI says dictation is now faster and more reliable. The company did not, however, give any figures for the speed or accuracy improvements, AI Times reported. South Korea’s Digital Today noted that OpenAI has not disclosed the new maximum recording length or whether dictation is now unlimited.
The rules for handling audio remain unchanged. OpenAI’s help documentation states that dictated audio is sent to the company’s models to be converted into text, AI Times reported. The audio is retained for as long as the conversation remains in the chat history. In the data controls settings, users can turn off the option to include audio recordings. Audio from dictation and voice conversations is then not shared for model training. The toggle does not, however, affect how audio is processed and retained for dictation.
The feature is coming to all ChatGPT plans worldwide: in the iOS and Android apps, on the web and in the desktop app. OpenAI has not announced a separate charge for dictation, South Korea’s Outsourcing Times reported, and besides the free tier the rollout covers the Go, Plus and Pro plans, among others.