gizmodo.com web signal

OpenAI launches ChatGPT Live Dictation, drops 10-minute cap

TL;DR

  • ChatGPT Live Dictation began rolling out November 18, 2026 across iOS, Android, web and desktop to every plan, including the free tier.
  • The feature transcribes speech in real time and removes the previous ten-minute recording limit, letting users dictate continuously.
  • OpenAI says consumer audio can be used for model training when the 'Improve the model for everyone' setting is left on; business outputs are not by default.

OpenAI started rolling out Live Dictation in ChatGPT on November 18, transcribing speech as users talk and scrapping the previous ten-minute recording cap. Gizmodo reports the feature is going out on iOS, Android, web and the desktop app, and is available on every ChatGPT plan, including the free tier.

OpenAI frames the change as part of a push to make AI interaction feel more natural, saying the goal is experiences "where talking to AI actually starts to feel like a real conversation." Users tap the microphone in the message box, watch the transcript build via a waveform they can show or hide, and edit before sending.

The quieter part of the launch is the data default. OpenAI states that "if you are a consumer user and have chosen to share audio to improve our models, we may train on audio from dictation," controllable by turning off "Improve the model for everyone" in Settings; business users' outputs are not shared for training by default. Gizmodo cites research that "participants' trust towards AI went up by an average of 33% when using voice commands compared to text prompts," and warns that pairing that trust uplift with an uncapped recording window may encourage vulnerable users to volunteer more than they would in text. The release lands amid a steady run of OpenAI product news on our AI Assistants tracker, including Google turning on Gemini Computer Use for Mac paid plans the same day.