Google Meet expands AI speech translation to 63 languages: what beginners should know

Announcement date: October 22, 2026

Google announced an expansion of AI-powered speech translation in Google Meet on October 22, 2026. The company says the feature supports 63 languages and lets individual participants choose the language they want to hear.

The update also introduces automatic detection of the language being spoken. Instead of setting one language pair for an entire meeting, participants can use their own translation controls. That could make a multilingual call feel less like a technical exercise and more like a conversation.

Speech translation is different from captions

Captions turn spoken words into text on a screen. Translation changes the language. Speech translation goes a step further by presenting translated audio for the listener.

For a beginner, separating those three ideas makes the settings easier to understand. You might want text in the original language, translated text, translated audio or a combination. Check which option is actually active rather than assuming a translation button covers every situation.

What Google changed

Google says the update uses Gemini 3.5 Live Translate, supports web and mobile devices, and offers listening support on Meet hardware. It describes automatically translated captions that match the translated audio, plus controls that let participants avoid translating languages they already understand.

Access and rollout conditions still matter. A feature announcement does not mean every account has every option immediately. Check the original announcement and the controls available in your own meeting before organising a call around the feature.

A useful example

Imagine a small volunteer group planning an event. Several people understand the discussion language, while one participant would follow more comfortably in another language. Individual controls could let that person use translation without changing the experience for everyone else.

The group should still make the important details easy to verify. Put the date, start time, address and names in the chat. A written reference gives people something to check if a spoken phrase was unclear.

Prepare the call, not just the button

Run a short practice call with the people involved. Ask each person whether they can hear the translation and find the captions. Agree on a simple way to pause the discussion if someone loses the thread.

Use short turns and leave room for questions. If a name, technical term or number matters, repeat it or type it. These habits help a group communicate clearly even when no translation tool is involved.

Do not assume a fluent-sounding translation preserves every nuance. A joke, a local expression or an ambiguous instruction can be difficult to translate. Ask the other person to confirm the intended meaning when the decision matters.

QueryFun’s take

The interesting change is the move towards a different language experience for each listener. That could make mixed-language meetings easier to join. The sensible first step is a short test with real participants, supported by written details and a willingness to slow down when something is unclear.

This report explains Google’s announcement. QueryFun has not independently tested translation quality across the supported languages.

Original source: Google Workspace’s speech-translation announcement. Company statements are attributed to their source. Practical examples and commentary are QueryFun’s explanation, not an independent product test.