Google's new AI transcription automatically edits out verbal fillers from speech

Google has updated its Gemini Audio system with new Gemini 3.5 models, introducing advanced transcription capabilities designed to make voice-to-text conversion cleaner and more precise. The system can now automatically detect and remove verbal fillers and hesitation sounds from recorded speech.

The release includes Gemini 3.5 Live, Gemini 3.5 Live Experimental, and Gemini 3.5 Transcribe. These models are engineered to enhance the precision of Google's voice-controlled AI features, ensuring reliable performance even in environments with background noise or irregular speech patterns.

Additionally, the new models introduce features for automatically recognizing specialized jargon and supporting more than 85 languages. This broadens the AI's capability to accurately process complex terminology across various professional domains.

For the global tech audience, these enhancements represent a significant step forward in natural language processing and speech recognition technology. As voice assistants and automated transcription tools become more sophisticated, the overall user experience in cross-lingual communication continues to improve.

According to The Verge, these updates highlight Google's ongoing commitment to refining its voice-controlled AI ecosystem, delivering smoother and more accurate interactions for everyday users.