Google Images just turned 25 years old. To mark the milestone, Google is rolling out two big updates: it is adding AI-powered image creation to AI Overviews, and it is giving the Google Images homepage a fresh, modern redesign. Both changes are rolling out over the next few weeks.
Let’s break down what’s changing, why it matters, and what it means for everyday searchers and for websites that depend on search traffic.
What’s New: Image Generation Inside AI Overviews
Until now, AI Overviews (the AI-written summaries that appear at the top of many Google search results) could only show pictures pulled from existing websites. That is changing.
Soon, when you search for something and Google can’t find the “perfect” existing image on the web, it will simply create one for you. Type a request, and AI Overviews will generate a brand-new, custom image built specifically to answer your query — no need to leave the search page.
This feature runs on Google’s own image model, known as Nano Banana, which the company has been rolling out across Search, Chrome, and its Gemini app throughout the year. One handy use case Google highlighted: asking Search to create a side-by-side visual comparison of two things, so you can see the differences at a glance instead of piecing it together from separate articles.
The rollout starts in English, and only in regions where AI Mode already supports image creation. Google hasn’t given a firm date for a full, worldwide launch — it’s a gradual rollout, and how far it spreads will depend on how AI Mode’s image tools expand more broadly.
Also read this blog: AI Overviews vs Featured Snippets — Key Differences
What’s New: A Redesigned Google Images Homepage
The second update is a complete makeover of Google Images itself. Instead of the plain search box most people remember, Google Images is becoming a browsable, real-time gallery.
Here’s what that means in practice:
- A live, scrolling feed of images pulled from across the web, refreshed continuously.
- Personalization — if you’re signed into your Google account, the images you see are tailored to your interests.
- Collections you can revisit — when you save an image, it gets added to a collection, and these collections show up as tabs right above the main gallery so you can jump back into a topic later.
- A simpler saved-items view — a Collections tab and an All Results tab let you switch between what you’ve saved and everything you’ve searched.
This redesign is starting on desktop in the United States, in English, before it likely expands further.
Why Is Google Doing This Now?
Google is tying these launches to the 25th anniversary of Google Images. The company has told a fun origin story about it: Google Images reportedly grew out of the massive spike in searches for Jennifer Lopez’s green Versace dress after she wore it to the Grammy Awards back in 2000. People wanted to see the dress, not just read text describing it — and that demand for visual search results never really went away.
Since then, Google has layered on tool after tool to make search more visual: Google Lens, “search with a photo,” Multisearch (combining a photo and text in one search), and Circle to Search on Android phones, which Google says is now available on more than 580 million Android devices. AI Overviews with generated images and the new Images homepage are the next step in that same visual-first direction.
How This Fits Into Google’s Bigger AI Picture
This isn’t happening in isolation. Nano Banana, the model doing the image generation, is the same technology Google has been pushing across its Gemini app for photo editing, style transfers, and combining multiple images into one. Google also recently introduced Google Pics, a separate, more advanced AI image creation and editing tool for Google Workspace aimed at professional-style editing — things like selecting and moving individual objects in a picture, or editing text within an image while keeping the original font style.
Put together, it’s clear Google wants AI-generated imagery to be available everywhere a user might need it — search results, chat conversations, and workplace documents alike.
What This Means for Searchers
For regular users, the upside is convenience: instead of scrolling through image results hoping to find exactly the picture you have in mind, you can just describe it and get something built to match. That’s especially useful for abstract ideas, comparisons, or concepts that simply don’t exist as a photo anywhere online.
Google has said, consistent with its policies elsewhere, that AI-generated images carry watermarking (including an invisible SynthID marker) so they can be identified as AI-made rather than a real photograph.
What This Means for Websites and SEO
For website owners, publishers, and marketers, this update deserves attention. AI Overviews already answer a lot of search queries directly on the results page, meaning users often don’t need to click through to a website at all. Now that AI Overviews can generate their own images instead of only pulling from web pages, that surface becomes even more self-contained — it no longer has to rely on finding a suitable picture already published somewhere on the internet.
In practical terms, this could mean:
- Fewer clicks on some image-driven searches, since Google can now produce a satisfying visual answer on its own.
- Less advantage for sites that used to win visibility simply by having a decent, relevant image, if AI Overviews just generates its own picture instead of sourcing one.
- More importance on original, high-quality content and images that genuinely can’t be replicated by an AI model, since generic or stock-style visuals will be easier for Google to recreate on demand.
The Bottom Line
Google is treating its 25th anniversary of Google Images as a launchpad for the next phase of visual search: an AI that doesn’t just find pictures, but creates them on request, wrapped inside a redesigned, personalized Images experience. Both features are rolling out gradually, starting in English in select regions, so it may be a few weeks before everyone sees them. But the direction is clear — Google wants search to be as visual as it is textual, and it’s betting its own AI models can fill in the gaps the open web can’t.