ChatGPT Live Dictation Now Shows Text as You Speak, With No 10-Minute Cap

ChatGPT Live Dictation: ChatGPT Live Dictation now displays your words on screen while you speak and no longer cuts off recordings after 10 minutes. OpenAI says the feature is rolling out to all plans, including free accounts, though the rollout is gradual. Here is what users should know about how their audio is stored and used for training.

ChatGPT Live Dictation, Digimart Global: OpenAI has updated dictation in ChatGPT. According to release notes dated November 18, words now appear on screen while the user speaks, rather than only after the recording ends.

The company has also removed the previous 10-minute limit on a single recording.

OpenAI said Live Dictation “is rolling out globally to all ChatGPT plans on iOS, Android, web, and the desktop app,” including free accounts. Because the rollout is gradual, some users may not see the feature yet.

The feature is the microphone icon in the message box, not voice mode, where ChatGPT speaks back. Tapping or clicking the waveform shows or hides the live text.

When recording stops, the transcript appears in the message box and can be edited before sending. Reporting by 9to5Mac noted that the old cap cut off longer recordings, which made dictation awkward for notes, long prompts and emails.

How dictated audio is stored and used

Recordings are sent to OpenAI and transcribed there. According to OpenAI’s dictation FAQ, audio is kept for as long as the chat remains in the user’s history.

If a chat is deleted, OpenAI deletes the audio clip within 30 days, except for clips it must keep for security or legal reasons and clips previously shared for training that were already separated from the account.

OpenAI says it trains on dictation audio only if the user has chosen to share audio to improve its models. The FAQ states: “If you are a consumer user and have chosen to share audio to improve our models, we may train on audio from dictation.” The ChatGPT Voice help page says the same.

The transcribed text is handled as a normal message, so the general “Improve the model for everyone” setting applies to it.

OpenAI’s data controls page says data sharing is enabled by default for personal Free, Plus and Pro accounts. Gizmodo noted that business accounts are excluded by default.

Users who dictate often and keep default settings may therefore have a significant amount of dictated text included in training data.

Privacy settings users can change

Both controls are under Settings, in Data controls. “Include your audio recordings” stops audio sharing only, covering both dictation and voice chat.

Turning off “Improve the model for everyone” also keeps chats out of training, and chat history stays where it is. Deleting a chat also deletes its audio within 30 days.

Daniel Brooks

Daniel Brooks covers the U.S. economy, financial markets, corporate developments, and business trends. His reporting makes complex economic stories understandable for everyday readers.

Leave a Reply

Your email address will not be published. Required fields are marked *

Sports

Explore Digimart Global