AI for everyone in every language

Google published an account of how it is extending AI language support beyond text translation, saying its technologies and products now power everyday interactions in more than 300 languages spoken by more than 7 billion people, which it describes as 86% of the global population. The company said its goal is to support the world's 1,000 most-spoken languages.

On speech, Google said Gemini 3.5 Live Translate handles real-time spoken translation across 70 languages and more than 2,000 language pairs, capturing code-switching and emotional cues. Gemini 3.5 Transcribe, described as its most precise speech-to-text model, converts raw audio into formatted text in noisy environments and with complex jargon, and powers the Rambler feature on Android Gboard. Google said its Universal Speech Model was trained on 12 million hours of audio and uses cross-lingual transfer learning, applying patterns from data-rich languages to under-resourced ones. It cited 25 years of open research and more than 400 peer-reviewed speech papers.

Google described three open-data partnerships: WAXAL, covering 27 Sub-Saharan African languages spoken by more than 100 million people across more than 26 countries; Project Vaani with the Indian Institute of Science and Bhashini, which has collected more than 30,000 hours of speech across 109 languages from over 155,000 speakers; and the Amplify Initiative, with more than 1,600 local experts and 20 universities contributing 15,000 multimodal data points. It also launched Language Explorer, which visualizes LinguaMeta, a repository mapping more than 7,000 spoken, written and signed languages.

For connectivity constraints, Google said more than 3 billion people lack reliable internet access, and it developed TranslateGemma, lightweight open translation models built from Gemini and trained across 55 languages that run on-device without cloud or internet. It supports Viamo's Ask Viamo Anything voice assistant on standard feature phones, piloted in Rwanda, which has used Gemini to answer more than 2 million questions. For accessibility, Sign Language-to-Text was trained across more than 50 sign languages and powers sign-to-text dictation in Gboard and Live Transcribe on Pixel 11, starting with American Sign Language to English, for the 70 million people worldwide who rely on sign language.

Why it matters: developers and product teams gain open translation models that work offline on-device and open speech datasets spanning underrepresented languages, while Google says its language technologies already connect more than five billion people across nine platforms including Search, Android, Chrome, YouTube and Google Play.