Static

ChatGPT can transcribe as you speak now with Live Dictation

First reported by The Verge ·

The signal ●○○○ Compiled by AI from The Verge, the single source so far
Why you might care

Your voice prompts for ChatGPT are now processed instantly, reducing the friction to start a conversation or task.

What happened

OpenAI has introduced a new Live Dictation feature for ChatGPT, allowing real-time transcription of audio prompts directly from the user's microphone. This update enables the AI to process spoken input as it's being delivered, eliminating the previous need to complete a recording before transcription. The feature is rolling out globally across all ChatGPT plans, including web, desktop, iOS, and Android applications. Users can view the live transcript and edit it in the prompt box. OpenAI has also removed the 10-minute recording limit for voice dictation and stated the feature has become more reliable, following previous voice mode enhancements and a beta 'record mode' for meeting transcriptions.

What it means

The integration of live dictation marks a significant step towards more seamless human-AI interaction, reducing latency and making voice a more fluid input method. By enabling real-time transcription and removing recording limits, OpenAI is lowering the barrier for using voice commands and dictation in daily tasks, potentially increasing user engagement and the practical application of ChatGPT beyond text-based queries.

This evolution of voice input in ChatGPT signals a broader trend towards more natural and intuitive interfaces in AI assistants, aiming to compete with dedicated dictation software and voice-controlled devices. The enhanced reliability and accessibility across all plans suggest a strategic push to solidify ChatGPT's position as a versatile tool for both casual users and professionals requiring efficient content generation and transcription.

AI-written summary. May contain errors.