Google's AI Transcription Edits Out 'Um' and 'Ah'
Google has rolled out an update to its Gemini Audio technology, introducing new transcription capabilities that can automatically detect specialized jargon and over 85 languages. The update includes a completely new addition called Gemini 3.5 Transcribe, which represents a major advancement from the previous Chirp 3 model. This new model allows users to edit transcriptions naturally with their voice and can automatically remove filler words like 'um' and 'uh.'
The Gemini 3.5 Transcribe model can also attribute speech for up to three speakers in pre-recorded audio, as well as provide word-level timestamps. Users can customize the vocabulary of the model to adapt to unique spelling requirements and specialized jargon. The update is rolling out starting today in English for all macOS Gemini app users.
Other updates include Gemini 3.5 Live and 3.5 Live Experimental, which build on existing speech recognition tech that powers Gemini's voice chat mode. Gemini 3.5 Live is better at handling mid-sentence interruptions, language recognition, and live visual processing, while Gemini 3.5 Live Experimental goes further by narrating its progress step by step in real time.