Google will debut three new AI experiments designed to help users learn foreign languages. These tools utilize Google’s Gemini large language model to identify objects and situations in the user’s direct environment and provide translations that can help users seek help or trigger conversations.
If you want to try out new experiments, you can find them on the Google Labs page. Google experiments are not apps, which means you don’t have to download anything to get started. You just need to click on the experiment you want to try to enter in the prompt.
Read more: Best AI Chatbots of 2025
When debuting these new features, Google will give a direct introduction to other foreign language learning services, which also focus on AI tools. Duolingo’s CEO recently announced that the company “will do it first” and Openai’s Chatgpt has the ability to start new foreign language conversations at any time as required.
Lessons: Describe the situation
Google’s new small course tool allows users to describe situations in which they learn vocabulary and grammar, which can help locals describe problems. Using the provided context, the tool will provide advice to help users understand how specific phrases tailored to their current issues to understand how to seek help.
Voice: Speak casually
Language hanging tools facilitate casual conversations about strict sentence structures and grammatical protocols, teaching users how to give up formalities and adapt to a more spoken foreign language approach. The language hang simulates conversations between native speakers and allows users to discover the meaning of any word or phrase in a series of messages. AI models sometimes recognize or hallucinate mistakenly, so when using this feature you need to double-check with another source.
Word Cam: Detect items in photos
The third and last new tool, Word Cam, uses Gemini to detect objects in the photos you take – providing translation for your surroundings in the foreign language you are learning. This feature helps you describe the world around you, but Gemini may not be able to accurately mark every subject you photograph. When using Word Cam, it’s still worth double-checking your translations with another source.
Google representative Maggie Shiels told CNET that the purpose of creating the language learning experiment is to “inspire developers who use Gemini to build different use cases and experiences.”
This particular set of experiments aims to use multimodal LLM as a way to promote bite-sized courses.
Google’s new features aren’t launched for every language – at least not yet. Small courses, language hanging and word cam currently support translations in Arabic, Chinese, French, German, Greek, Greek, Hebrew, Hindi, Italian, Japanese, Korean, Portuguese, Russian, Russian, Spanish and Turkish.
Shiels says that like other Google Labs experiments, Tiny Thine lessons, s-language and Word Cam – are not products, nor are they permanent features.
“It’s a limited time tool that will eventually be sunset,” she told CNET. “We want developers to have fun.”
