During Google I/O 2025, the company announced the rollout of real-time speech translation to Google Meet.
This recently launched feature uses a large language audio model from Google DeepMind to enable a natural conversation with anyone in a different language.
Real-time speech translation
Speech translation in Meet is able to translate spoken words into the listener’s preferred language in real time.
The company further said that the latest feature is packed with a range of use cases. For example, users can use it to have English-speaking grandchildren having a conversation with their Spanish-speaking grandparents.
Notably, this useful tool may help companies that operate worldwide, enabling global colleagues to connect and chat in real-time.
However, the latency for speech translation is too low, enabling several people to chat together, which has remained impossible until now.
Whenever the person on the other side speaks, users faintly listen to their original voice, along with the translated speech overlaid on top.
Availability
Speech translation feature in Google Meet will be available to consumer AI subscribers in beta from Tuesday.
To note, the feature will be accessible in several languages, including English, Spanish, and more.
Google further stated that it’s currently creating the speech translation in Meet for businesses, with initial testing launching to Workspace customers this year.
