Google Meet Introduces Real-Time AI Voice Translation With Hebrew Support
Google Meet has integrated real-time bidirectional voice translation supporting 63 languages, including Hebrew. Powered by Gemini AI, the feature translates speech and attempts to mimic speakers' voices.
Google has expanded its real-time voice translation feature in Google Meet to support 63 languages, including Hebrew. Powered by the Gemini AI model, the tool provides bidirectional translation capabilities during video conferences.
How Live Translation Works in Google Meet
Users can activate the feature by clicking the Google Translate-style icon within the interface. The system allows participants to select their preferred listening language while also specifying languages that should remain untranslated. Supported languages now include Arabic, Amharic, French, German, English, Persian, Russian, Hebrew, and dozens of others.
Testing shows that when the system detects speech requiring translation, it automatically lowers the original speaker's volume. A translated dubbed voice then plays with a delay of one to two seconds.
Voice Cloning and Future Improvements
Beyond basic translation, the underlying Gemini technology attempts to mimic the original speaker's voice characteristics, adjusting dynamically when different participants join the conversation. While the voice cloning feature occasionally exhibits minor inconsistencies or shifts between speakers, the overall performance marks a significant step forward in breaking down global language barriers.
