{
  "id": 7581979,
  "title": "AI for everyone in every language",
  "url": "https://urgent.news/2026/09/15/ai-for-everyone-in-every-language",
  "topic": "ai",
  "section": "AI",
  "published": "2026-09-15T16:00:00.000Z",
  "source": {
    "name": "Google AI Blog",
    "slug": "google-ai-blog",
    "url": "https://blog.google/innovation-and-ai/technology/ai/ai-for-every-language/"
  },
  "original_language": "en",
  "account": "Google is leveraging artificial intelligence to facilitate communication across hundreds of languages, including those that have been historically underrepresented in technology. The company aims to understand the subtleties of human communication, such as emotion, tone, and slang, which go beyond simple text translation. To achieve this, Google is collaborating with local communities worldwide to ensure its tools cater to everyone, irrespective of their internet access.\n\nToday, Google's technologies and products support over 300 languages, spoken by more than 7 billion people, constituting 86% of the global population. Achieving this milestone is a significant accomplishment, yet it highlights the ongoing need to address the digital representation gap for thousands of living languages and dialects. Initially, Google Translate was launched in 2006 with the primary goal of breaking down language barriers. With the advancements in AI, the vision has expanded from a few languages to over 250 today.\n\nGoogle's research and development focus on creating systems that respect cultural nuances and the richness of human language to ensure everyone can participate and be understood. Traditional speech recognition systems followed a linear process of transcribing audio into text, processing that text, and synthesizing it back into audio. However, this approach failed to capture the essence of human communication, which includes tone, pacing, emotion, and context. Recognizing this limitation, Google transitioned from text transcripts to native audio intelligence, training models like Gemini to process audio directly, capturing both sound and intent.\n\nSince the web predominantly features a few dominant languages, Google had to rethink data gathering methods. Local grassroots partnerships were established to collect data more effectively. Notable collaborations include initiatives to enhance language data representation, such as the Language Explorer tool, a platform that visualizes LinguaMeta, the world's largest open-source language data repository. This interactive tool maps over 7,000 spoken, written, and signed languages, recognized for its design innovation by Fast Company.\n\nDespite the progress, over 3 billion people still lack reliable internet access. To make technology accessible to everyone, including those in low-resource regions with limited or intermittent connectivity, Google developed TranslateGemma. This family of lightweight open translation models, built from Gemini and trained across 55 languages, operates efficiently on-device, eliminating the need for cloud or internet connections.\n\nRunning AI models requires advanced hardware, which excludes many people in feature phone regions. To bridge this gap, Google supports organizations like Viamo to create \"Ask Viamo Anything\" (AVA), a voice AI assistant that brings Gemini's capabilities to standard feature phones. In Rwanda, AVA has been successfully piloted, answering over 2 million questions using Gemini's technology.\n\nLanguage encompasses various forms of communication, including non-standard speech and sign language. Google is addressing these aspects by designing accessible solutions, such as Sign Language-to-Text (SL2T). Trained across 50+ sign languages, SL2T powers sign-to-text dictation in Gboard and Live Transcribe on Pixel 11, starting with American Sign Language (ASL) to English. This initiative is a crucial step towards ensuring technology accessibility for the 70 million people worldwide who rely on sign language.\n\nUnderstanding cultural context is vital in developing language technology. Google collaborates directly with local communities to improve cultural accuracy in tools. For instance, in New Zealand, they worked with Māori language experts to enhance place name pronunciations in Google Maps, making the experience more precise and culturally relevant. This approach ensures that technology reflects the language and heritage of the communities it serves. Over the years, Google has learned that technology should not limit the spectrum of human expression but rather expand it.",
  "summary": "We’re moving beyond traditional text translation to build models that understand the world’s rich, living languages exactly as they are expressed.",
  "key_points": [
    "Google uses AI to support over 300 languages for communication",
    "Transitioned from text transcripts to native audio intelligence",
    "Developed TranslateGemma for low-resource regions with limited connectivity"
  ],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 2,
    "also_reported_by": [
      {
        "outlet": "Google Blog",
        "title": "AI for everyone in every language",
        "url": "https://urgent.news/2026/09/15/ai-for-everyone-in-every-language-7585401",
        "published": "2026-09-15T16:00:00.000Z"
      }
    ]
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}