Google's New AI Transcription Cleans Up Your Speech: 'Ums' and 'Ahs' Removed

Google has introduced new Gemini Audio models designed to enhance AI transcription by automatically removing filler words like "ums" and "ahs" from recorded speech. This update arrives as users eagerly await the release of the delayed Gemini 3.5 Pro, which remains unlaunched as of early 2026. The new audio models aim to produce cleaner, more polished transcripts by filtering out common verbal hesitations, making AI-generated text more natural and professional. This development reflects a broader industry trend toward refining speech-to-text technology, particularly in areas such as meeting notes, content creation, and accessibility tools. While no release date for Gemini 3.5 Pro has been confirmed, the introduction of these audio models suggests that Google is actively enhancing its AI offerings, possibly setting the stage for a more comprehensive update. As the demand for efficient and accurate transcription tools grows, this feature could give Google a competitive edge in the evolving AI landscape.

via The Verge AI