Skip to content
IFAI

AI tools

Google Translate taps Gemini for live speech and language coaching

Published Reach 48

Google just added two major AI-powered features to its Translate app, using Gemini models for real-time live conversations across 70+ languages and custom interactive language practice. The live mode auto-detects pauses and filters out cafe noise with a single tap, while the practice feature builds personalized speaking scenarios. Live translation launches today in the US, India, and Mexico, with learning tools in beta.

Google Translate taps Gemini for live speech and language coaching
Image: blog.google

If you have ever awkwardly held your phone between yourself and a waiter in a noisy foreign restaurant, taking turns tapping tiny language-swap buttons, Google feels your pain.

On Tuesday, Google announced a major AI upgrade to Google Translate, bringing two new capabilities powered by its Gemini models: a fluid, hands-free live conversation mode and a personalized language learning practice feature.

To put Google's language footprint into perspective, people translate roughly 1 trillion words every month across Translate, Google Search, Lens, and Circle to Search.

One tap for 70 languages

The new Live Translate experience lets two people hold a back-and-forth conversation across more than 70 languages, including Spanish, French, Hindi, Korean, Arabic, and Tamil.

Instead of making users manually tap back and forth to switch who is speaking, Gemini listens continuously. It automatically detects natural conversation pauses, regional accents, and intonations, smoothly switching the translation direction while providing on-screen transcripts and spoken audio.

Google also loaded in voice recognition models trained on sound isolation, designed to keep conversations intelligible even inside busy airports or noisy cafes. The live conversation tool is rolling out first to Android and iOS users in the U.S., India, and Mexico.

Google comes for the green owl

The bigger surprise is Translate's new interactive language practice mode, which TechCrunch noted takes aim straight at Duolingo.

Users can tap Practice inside the app, pick their proficiency level, and set personal goals. Gemini then generates customized, real-time conversation scenarios on the fly.

Learners can either listen to spoken dialogues and tap the words they hear to build listening comprehension, or practice speaking out loud with adaptive hints when stuck. The tool tracks daily progress and was developed alongside language acquisition experts.

The beta feature is landing this week on Android and iOS. It currently supports English speakers practicing Spanish and French, as well as Spanish, French, and Portuguese speakers practicing English.

Why it matters

For everyday travelers and learners, Google Translate is shifting from an emergency survival utility into an active, on-demand tutor. You no longer have to juggle a dictionary app alongside a gamified flashcard app just to prepare for a weekend abroad.

For the edtech industry, this is an existential warning shot. Platforms like Duolingo built massive businesses on scripted, gamified paths. But when an existing utility with a billion users can generate infinite, hyper-personalized spoken roleplay scenarios for free, proprietary language apps will need to justify why users should pay for rigid lesson trees.

We think embedding Gemini directly into Translate's classic interface is a brilliant distribution play. The biggest hurdle will be pedagogical accuracy: large models can be overly forgiving or hallucinate idioms. If Google can maintain natural conversational flow without leading learners astray, Translate could redefine how people pick up a second language.

Sources (2)

  1. [1]
    blog.google(opens in a new tab)Originalblog.google
  2. [2]
    techcrunch.com(opens in a new tab)Reporttechcrunch.com

Best of AI, daily