Google Images turns 25 with AI generation right in Search

Twenty-five years ago, Jennifer Lopez walked a red carpet in an iconic green Versace dress, and millions rushed to Google only to find a wall of blue text links. That collective frustration gave birth to Google Images in July 2001.

To mark its 25th anniversary, Google is rolling out two major updates: a personalized, browseable visual homepage and native image generation baked straight into Search.

Generate what the web doesn't have

Sometimes the visual you need simply does not exist online. Google is tackling that gap by integrating its latest Nano Banana model directly into AI Overviews in Search.

Instead of hunting through irrelevant uploads, you can type a plain text prompt and get a custom, high-quality image generated on the fly. The rollout will begin over the coming weeks in English across regions where AI Mode image creation is already supported.

Google Images on desktop is also getting its biggest interface shakeup in years. The blank search bar is being replaced with an immersive, real-time gallery tailored to your personal interests.

As you discover and save images into collections, they automatically appear as tabs above the main feed. This personalized discovery home will launch first for signed-in desktop users in the U.S. in English.

25 years of teaching search to see

Visual search has traveled a long road from matching file names to understanding camera feeds.

Google rolled out Similar Images in 2009, introduced reverse image lookup in 2011, and brought real-time object scanning to phones with Google Lens in 2018. Recent years shifted the focus toward multimodal reasoning, from combining text and photos in Multisearch to letting users circle items on their screens.

Google noted that Circle to Search is now active on over 580 million Android devices globally. The feature recently gained multi-object recognition, breaking down complex scenes into dozens of parallel sub-queries.

With users now able to drop multiple photos into the search box for reasoning, visual search has shifted entirely from keyword matching to ambient computing.

Why it matters

We think folding the Nano Banana generator straight into AI Overviews is a clever move that collapses the boundary between searching and creating.

For everyday users and casual creators, it removes the friction of jumping to standalone image platforms like Midjourney when a web query turns up empty. If reality does not have the picture, Google simply makes it for you in the same tab.

Turning the desktop homepage into an algorithmic discovery feed shows Google taking dead aim at Pinterest for visual inspiration and moodboarding. Google clearly wants you to browse passively rather than just query and bounce.

However, blending synthetic outputs seamlessly into core search results creates real headaches around visual provenance and misinformation. Competitors in the stock photography and standalone generative tool spaces should also be worried, as Google's default distribution remains nearly impossible to beat.