Google is debuting three new AI experiments that are intended to help users learn foreign languages on the go. The tools utilize Google’s Gemini large language model to identify objects and situations in a user’s immediate environment and provide translations that could help users ask for help or spark a conversation.

If you want to give the new experiments a try, you can find them on the Google Labs webpage. Google experiments aren’t applications, which means you don’t have to download anything to get started. You can just click into the experiment you want to try and begin typing in your prompts.

Read more: Best AI Chatbots of 2025

In debuting these new features, Google is going head-to-head with other foreign language-learning services that are also focusing on AI tools. Duolingo’s CEO recently announced that the company «will be going AI-first,» and OpenAI’s ChatGPT has the ability to begin new foreign-language conversations at any time upon request.

Tiny Lesson: Describe a situation

Google’s new Tiny Lesson tool allows users to describe a situation they’re in to learn vocabulary and grammar that can help describe a problem to the locals. Using the provided context, the tool will provide suggestions that aid users in understanding how to ask for help if they haven’t learned specific phrases tailored to their current issue.

Slang Hang: Casual talk

The Slang Hang tool promotes casual conversation over rigid sentence structure and grammatical agreement, teaching users how to drop the formalities and adapt a more colloquial way of speaking a foreign language. Slang Hang simulates conversations between native speakers and lets users discover what any words or phrases in the series of messages mean. The AI model sometimes misidentifies or hallucinates words, so you’ll need to double-check with another source when using this feature.

Word Cam: Detect items in photographs

The third and final new tool, Word Cam, uses Gemini to detect objects in photographs you take — providing you translations for your surroundings in the foreign language you’re learning. This feature helps you describe the world around you, but it’s possible that Gemini may not accurately label every single object in a picture you take. It’s still worth double-checking the translations you’re provided against another source while using Word Cam.

The language-learning experiments were created as a way to «inspire developers using Gemini for building different use cases and experiences,» Google representative Maggie Shiels told CNET.

This particular set of experiments is meant to focus on using the multimodal LLM as a way to promote bite-sized lessons on the go.

Google’s new features aren’t launching for every language — at least, not yet. Tiny Lesson, Slang Hang and Word Cam currently support translations for the Arabic, Chinese, English, French, German, Greek, Hebrew, Hindi, Italian, Japanese, Korean, Portuguese, Russian, Spanish and Turkish languages.

Shiels said that Tiny Lesson, Slang Hang and Word Cam — like other Google Labs experiments — are not products and are not meant to be permanent features. 

«This is a limited-time tool that will eventually sunset,» she told CNET. «We hope that developers have fun playing around.»

As Google starts to revive its Google Glass concept, Meta is already a step ahead with new artificial intelligence functions coming to glasses this summer. The Ray-Ban smart glasses, in partnership with Meta, are getting several powerful AI updates for US and Canadian users. 

Operating the Meta View app on a connected smartphone, users of Ray-Ban smart glasses will also be able to use the «Hey Meta, start live AI» command to give Meta AI a live view of whatever they are seeing through their glasses. 

Similar to Google’s Gemini demo, users will be able to ask Meta AI conversational questions about what it sees and how it might solve problems. Meta provided the example of Meta AI giving possible substitutes for butter based on what it sees when you look in the pantry. 

Even without live AI, you’ll be able to ask specific questions about objects that you’re looking at.

In addition to new seasonal looks, Ray-Ban’s smart glasses also will be able to use the «Hey Meta, start live translation» command to automatically translate incoming languages including English, French, Italian and Spanish. The glasses’ speakers will translate as other people talk and you can hold up your phone so the other party can see a translated transcript too. 

Meta AI and concerns about being filmed

When I reached Inna Tokarev Sela, CEO and founder of AI data company illumex about privacy issues with smart glasses like these, she mentioned that in her own experience with Ray-Ban smart glasses, people usually reacted when they noticed the recording indicator light, which meant the glasses were watching. That can make some people uneasy, whether they are concerned about being filmed by a stranger or by what Meta may be doing with all that visual data it’s collecting.

«In the new models you can control the notification light, which could pose a privacy risk,» Sela said. «But everyone films everyone all the time anyway at touristy landmarks, public events, etc. What I expect is that Meta will not divulge any information on anyone, unless they register and explicitly give their consent.»

This could lead to other consent headaches too, depending on if users are recording for other purposes. «For example, users should be able to opt in and choose the type of information to expose when they’re in someone’s frame — similar to LinkedIn, for example,» Sela said. «Of course, any recording resulting from the glasses should not be admissible to use in a court of law, as with any other kind of recording, without explicit permission.»

Additional updates and rollout schedules

Along with the AI upgrades, Ray-Ban’s smart glasses will be able to post automatically on Instagram or send a message on Messenger with the right voice commands. New compatibility with music streaming services also will allow you to play songs through Amazon Music, Apple Music and Spotify on your glasses in lieu of earbuds.

Meta reports that the rollout of these new features will happen this spring and summer, along with object recognition updates for EU users arriving in late April and early May. 

Meta and Ray-Ban didn’t immediately respond to a request for further comment. 

The effect of tariffs can be felt on almost everything you’re buying right now, but how much are they actually costing you? It can be hard to keep track. Some companies have been open about price hikes, but Amazon will not be one of them.

A report from Punchbowl News on Tuesday cited a source familiar with the plan as saying that Amazon intends to make it known to shoppers how much of a difference the tariffs are making to the price of items they want to buy. Amazon later shut down the claim.

«The team that runs our ultra low-cost Amazon Haul store considered the idea of listing import charges on certain products,» company spokesperson Tim Doyle said in an emailed statement. «This was never approved and is not going to happen.»

Even before Amazon denied the plan, President Donald Trump made it clear he was unhappy with the idea of the company listing tariff costs. «This is a hostile and political act by Amazon,» White House press secretary Karoline Leavitt told reporters during a briefing.

Tariffs have been sending costs soaring across shopping platforms, with reports that products on Chinese sites Temu and Shein have seen price hikes of up to 377%. Even though Amazon is a US company, it’s also feeling the impacts of increased import charges, with many products available across the platform originating in China.

«This is another reason why Americans should buy American,» Leavitt said Tuesday. Separately on Tuesday, Amazon announced its plans for the upcoming Prime Day sale. The shopping event will once again take place in July and offer savings across all categories of products, though it’s hard to say whether the savings this time will be as significant as on previous Prime Day sales.