Google is turning Gemini into a place to explore a topic, not only ask about one. On October 28, 2026, Google announced that the Gemini app now has a library of immersive explorations: interactive, visual tours that let you "explore topics layer by layer, from microscopic systems to deep space." The first examples are tracing Silk Road trade routes, looking inside a jet engine and studying the physics of lightning.
It is a modest launch with a clear idea behind it: a good explanation is often a zoomable one. Here is what Google says, how to try it, and the questions the announcement leaves open.
TL;DR: immersive explorations at a glance
| Question | Answer from Google's post |
|---|---|
| What is it? | A library of interactive, visual tours inside the Gemini app |
| How does it work? | Tap a label to dig deeper, moving from large scales down to molecules and atoms |
| Examples | Silk Road trade routes, inside a jet engine, the physics of lightning |
| Where? | The Gemini student hub (Try immersive explorations) or gemini.google.com/immersive-explorations |
| Languages | English, Indonesian, Hindi, Brazilian Portuguese, French, simplified Chinese; English-only elsewhere for now |
| Coming | Audio narration for each layer; access directly inside a Gemini chat |
| Not stated | Pricing or plan limits, the underlying model, how content is checked for accuracy |
How do I open immersive explorations?
Go to the Gemini student hub in the Gemini app and click Try immersive explorations. The library is also at gemini.google.com/immersive-explorations. Google says the explorations will "soon" be available directly in a chat conversation with Gemini, which would let you ask a question and be handed a tour instead of a text answer.
Each exploration opens into a guided, interactive scene. Selecting a label reveals more detail and moves you down a level of scale, for instance from a planet to a molecule to an atom. Audio narration to guide you through each layer is listed as coming.
Why layers matter for learning
Layered explanations are a long-standing teaching device: start with the overview, then let the learner choose which part to open. A jet engine is a good example. At the top level you see the engine as a whole; one layer down, the compressor, combustor and turbine; one more down, the blades, their materials and the airflow around them. A learner who already knows the overview can skip it, while a beginner is not overwhelmed.
Compare that with a chat answer, which is linear and decided by the model. An exploration puts the learner in control of depth. That control is the educational bet. Google has not published any study on this product, so treat the benefit as plausible rather than proven.
What this adds to Gemini's student push
The feature sits in the student hub, which has been growing through the autumn. Related moves we have covered include Chrome's Gemini student features such as quizzes and tab sync, the Gemini Notebook usage limits with a five-hour reset, and the Gemini small business hub, which shows the same pattern of themed hubs inside the app. Reusable instructions now live in Gemini Skills, which replaced Gems.
Competitors are chasing the same group. OpenAI's report on how teens use ChatGPT, with a college planner and flashcards shows where usage is heading, and Anthropic is courting campuses through its Claude campus ambassador programme. Immersive explorations are Google's attempt to differentiate on visual, interactive explanation rather than on chat.
What Google has not said
The announcement is short. Several things a careful reader, parent or teacher would want to know are missing:
- The model and method. It does not say which Gemini model generates the explorations, whether scenes are generated on demand or pre-built, or how they are reviewed.
- Accuracy checks. Interactive diagrams of physics or biology can be wrong in subtle ways. Google does not describe expert review.
- Pricing and limits. The post says the library is available to all users in the listed languages and mentions no plan tiers.
- Age and school use. It does not say whether the feature works with school accounts or any age restrictions.
- Data. There is no statement on what interaction data is collected.
- Accessibility. Narration is promised, but keyboard and screen reader support is not described.
None of this means a problem exists. It means the announcement is a product teaser rather than documentation.
How to use it well
A few habits make an interactive explainer more useful than a flashy one.
- Predict before you drill. Before opening a layer, write down what you expect to see. The surprise is where learning happens.
- Cross-check one fact. Pick one claim from a layer, such as a number or a mechanism, and confirm it in a textbook or a primary source. If the check fails, tell Gemini and note the topic as unreliable.
- Ask a follow-up in chat once it ships. Because chat access is coming, plan to ask why a component works the way it does, rather than only clicking labels.
- Teach it back. Explain the exploration to someone else in two minutes. If you cannot, you looked rather than learned.
- For teachers: run one exploration in class, ask students to find one error or simplification, and discuss it.
What people are asking
Is this the same as asking Gemini to explain something? No. A chat answer is text the model chooses to write. An exploration is a visual scene with labels you choose to open. The experience is closer to an interactive diagram in a textbook than to a conversation, although chat access is promised.
Can I make my own exploration? Google's post describes a library, not a creation tool. It does not say you can request a custom topic, so for now assume you are choosing from what exists. If a topic you need is missing, Gemini's regular chat and its student features remain the fallback.
Is it only for students? The entry point is the student hub, but the post says the library is available to all users in the listed languages. A curious adult can use it the same way.
Can I trust the physics? Treat each exploration as a good first pass. Lightning and jet engines are well-studied topics, so errors are easier to spot, but a simplified diagram can still mislead. Compare against a trusted source before you rely on a detail for an exam or a job.
Will it work offline or on a school device? The post is silent on both. Schools that restrict consumer AI features should check their admin settings, and any school considering it should ask Google about data handling first.
Why a visual format fits the moment
Large language models are good at fluent text and weaker at showing structure. Interactive scenes hand the structure back to the reader: where things are, how big they are relative to each other, and what lies inside what. That matters most in subjects where scale is the hard idea, such as cell biology, astronomy and engineering. It is also why labels that jump between scales, planets to molecules, are a natural fit. If Google extends the library with narration and chat, the same format could suit history and geography, as the Silk Road example suggests, where the layers are routes, cities, goods and people rather than physical scale.
Languages and reach
Support for English, Indonesian, Hindi, Brazilian Portuguese, French and simplified Chinese is a deliberate spread across large student populations. Everyone else gets an English-only experience for now, with "more languages" promised and no dates. If you teach in another language, test whether a mixed-language experience helps or confuses your students before assigning it.
Where it fits among AI learning tools
Interactive explanation tools fall on a spectrum. At one end are chat tutors that answer questions in text. At the other are simulations built by educators, where every state is designed and tested. Immersive explorations sit in the middle: more structured than chat, probably less rigorously checked than a hand-built simulation. That is a reasonable place for a general-purpose assistant, as long as users understand the difference.
For builders interested in making their own interactive lessons with AI, our posts on agent skills and on Google's AI in science report, which discusses where AI explanations are and are not reliable, are good starting points.
Bottom line
Immersive explorations are a small, sensible step: zoomable, visual explanations in Gemini, open now in six languages, with chat access and narration to come. The unanswered questions about accuracy review, models and data are the ones to watch. Try a topic you already know well first; if the layers match what you know, you can trust it a little more on topics you do not.
Related reading
- Chrome's Gemini student features
- Gemini Notebook flexible usage limits
- Gemini Skills replace Gems
- Gemini small business hub
- ChatGPT for teens usage report
- Google's AI in science report
- Official: Google's announcement and the exploration library
Details are accurate as of October 28, 2026 and come from Google's announcement; availability and languages may change.