Dictate Natural Brain-Dumps Using Improved Speech-to-Text Accuracy
Capture multilingual ideas on-the-go with noise-robust dictation.
Take advantage of the improved speech-to-text engine in ChatGPT for hands-free brainstorming. It is specifically optimized for noisy environments and multilingual users.
The Scenario
You are walking through a noisy city or working in a busy office and need to capture a brain-dump of ideas that might involve technical terms in multiple languages.
Before & after
Recording ideas in a busy cafe or while commuting often leads to messy transcripts that require 15-20 minutes of manual editing to fix typos and language switches.
Dictate your notes directly into ChatGPT using the updated speech-to-text model. It captures mixed-language ideas and noisy backgrounds perfectly in 1-2 minutes.
The Prompt
[ACTIVATE DICTATION MODE] I am going to dictate a series of thoughts about [TOPIC]. Please transcribe this accurately, even if I switch between [LANGUAGE A] and [LANGUAGE B].
The June 2026 update significantly improved dictation accuracy for multilingual speakers and added robustness for noisy environments.
Source
Model Release Notes | OpenAI Help Center"The update improves transcription accuracy across languages and accents... Dictation is also more robust in noisy public spaces."