Google Meet introduces real-time speech translation feature

Google has announced the launch of real-time speech translation in Google Meet, aiming to enable seamless multilingual conversations.

The announcement came during Google I/O 2025, where the company unveiled the feature powered by a large language audio model developed by Google DeepMind.

The new capability translates spoken words into the listener’s preferred language in real time, while preserving the original speaker’s voice, tone, and expression.

Google says the technology enables fluid, natural dialogue between users who speak different languages, with minimal latency—an advancement that could significantly improve global communication.

When a participant speaks, their original voice can still be faintly heard underneath the translated audio.

The feature allows multiple users to participate in a conversation simultaneously, a development Google claims was not feasible until now.

Read more: UAE to introduce Artificial Intelligence as subject in all govt schools

The speech translation function will begin rolling out in beta to consumer AI subscribers starting Tuesday. Initially, translations will be available between English and Spanish, with additional languages including Italian, German, and Portuguese expected in the coming weeks.

Google also plans to expand the feature for business use, with early testing scheduled for Workspace customers later this year.

The company highlighted potential applications ranging from family communications across generations to multinational corporate meetings.