OpenAI Launches Live Dictation for ChatGPT

The feature transcribes speech in real time and removes the earlier ten-minute recording limit for all ChatGPT plans.

Updated on Nov. 19, 2026 in Artificial Intelligence

OpenAI Launches Live Dictation for ChatGPT

Live Poll

Do you trust AI chatbots with the personal details you share with them?

OpenAI has launched a Live Dictation feature for ChatGPT that transcribes speech as the user talks. The feature began rolling out Wednesday across iOS, Android, web and the desktop app for all ChatGPT plans, and it eliminates the ten-minute recording limit of the previous dictation tool.

Why it matters

The launch reflects OpenAI's broader push to make interacting with AI feel more natural than typing one text prompt after another. Users outside the company's business tier should note that dictation audio may be used to train models if they have chosen to share audio for that purpose.

The previous ChatGPT dictation tool generated a complete transcription only after the user finished speaking and was capped at ten minutes of recording. Live Dictation removes that cap and produces text as speech occurs, though the stub does not specify latency or accuracy figures.

The players

OpenAI

OpenAI is an artificial intelligence research and deployment company best known for developing the ChatGPT chatbot and the GPT family of language models.

The details

In a demo video published on X, an actor used Live Dictation to ramble about his relationship with a close friend and asked ChatGPT to organize the material into bullet points for a best man speech. OpenAI's website states that consumer users who opt in to sharing audio for model improvement may have their dictation audio used for training, while outputs from business users are not shared by default.

Timeline

  1. In July, OpenAI released the GPT-Live-1 voice model.

  2. In 2025, Jabra and the London School of Economics conducted a study of 171 knowledge workers on trust in voice interaction.

  3. On Wednesday, Live Dictation began rolling out to all ChatGPT plans.

The Tech Race

Live Dictation extends OpenAI's GPT-Live-1 voice model, released in July, by moving real-time speech input into the mainstream ChatGPT app. The move positions OpenAI to compete on voice interaction with rival assistants that are also pushing spoken input as a default.

Users on any ChatGPT plan can now dictate longer passages without hitting a ten-minute cap, which makes drafting notes or speeches by voice more practical. Anyone who does not want dictation audio used for training should turn off "Improve the model for everyone" in ChatGPT's Settings.

The takeaway

Speaking to an AI assistant may feel more natural than typing, and the 2025 study suggests that voice input can raise trust in the tool. If you dictate sensitive material, check your data-sharing setting first.

Further reading

For broader context on how voice is reshaping AI tools, see Artificial Intelligence coverage on Wisevoter.

Live Poll

Do you trust AI chatbots with the personal details you share with them?