Google Meet has introduced a new real-time spoken translation feature, developed in collaboration with Google DeepMind and Google Research.
The project, which initially was expected to take five years, came to life in just two.
The breakthrough allows people to communicate in their own voices across different languages, making meetings, travel planning, and personal calls more seamless.
Earlier translation systems required transcribing speech, translating text, and then converting it back to audio, a process that caused delays of 10 to 20 seconds and removed the speaker’s unique voice.
The new model processes audio directly, drastically reducing latency to just two or three seconds.
This timing closely mimics the way a human interpreter works, allowing for fluid, natural conversation across languages in real time.
To refine the feature, teams worked closely with linguists to capture nuances in accents, grammar, and idioms.
While challenges remain, such as literal translations or tricky phrases in structurally different languages like German, Google expects future updates with advanced language models to improve tone, irony, and contextual understanding.
By combining technical innovation with cross-team collaboration, AI has made global communication in Google Meet more accessible and human than ever before.
You may also want to check out some of our other recent updates.
Subscribe to Vavoza Insider to access the latest business and marketing insights, news, and trends daily with unmatched speed and conciseness! 🗞️
