Abishek D Praphullalumar
Google Meet is getting a major AI upgrade with the introduction of real-time speech translation, announced at Google I/O 2025. Powered by a large language audio model from DeepMind, the new feature enables natural, free-flowing conversations between speakers of different languages while preserving voice tone and expression.
The translation happens in real time, with very low latency, allowing multiple participants to chat seamlessly. Listeners will hear the translated voice overlaid on the original, which remains faintly audible in the background. Google says it’s ideal for everything from multilingual family chats to global business meetings.
The feature begins rolling out in beta today for consumer AI subscribers, starting with English and Spanish. Support for additional languages, including Italian, German, and Portuguese, will arrive in the coming weeks. Google also confirmed that Workspace customers will get access later this year as part of a phased business rollout.
