Google is celebrating 25 years of Google Images by introducing two significant AI-powered features: a browsable homepage with organized collections and direct image generation within search results, powered by its latest Nano Banana model. Meanwhile, the company faces a new copyright infringement lawsuit from major publishers and authors who claim Google used their works to train Gemini without permission or compensation.

What New Features Is Google Adding to Google Images?

Google is rolling out a redesigned Google Images homepage featuring a dynamic, real-time gallery tailored to user interests. The most practical addition is a tabbed collections system that works similarly to Pinterest boards or TikTok collections. Users can save thematically related images to organized tabs for easy access later.

For example, someone researching a trip to Malaysia could create a collection titled "Malaysia" and save all their travel inspiration in one place. This feature will be available over the coming weeks on desktop in the United States in English.

The second feature brings image generation directly into AI Overviews in Search. Using text prompts powered by Gemini's Nano Banana model, users can generate custom visuals to help with decision-making. If you're trying to decide between paint colors for a bedroom, you could type "create a visual that compares a master bedroom with yellow-gold walls versus sage green walls," and the AI would generate a comparison image.

"Especially with these AI models, we have an ability to do way more to get towards more of this natural and intuitive expression of a question. I think that's where we're more and more anchored: How do you let people express these questions really intuitively and then obviously answer them to the best that we can?" said Lou Wang, Google Lens co-founder.

Lou Wang, Google Lens co-founder

This image generation feature will arrive over the coming weeks in regions that currently support image creation in AI Mode, available in English.

How Has Google's Visual Search Evolved Over Two Decades?

• 2001 Launch: Google Images debuted after Jennifer Lopez's iconic green Versace dress at the 2000 Grammy Awards sparked massive internet interest in finding similar images.
• Reverse Image Search: Over the following decade, the tool evolved from text-based image queries to reverse-image searches using uploaded images or URLs.
• Google Lens Integration: In 2017, Google introduced Google Lens, initially integrated with Google Photos and Google Assistant, then moved to the main search box.
• Multisearch and Circle to Search: Google added Multisearch, which combines text and image inputs, and Circle to Search, enabling context-aware image search without switching apps.
• Recent AI Enhancements: Over the past two years, AI Mode integrations, Search Live capabilities, visual reasoning powered by Gemini models, "Find the Look" fashion outfit matching, and multi-image search inputs have made searching with imagery more natural and intuitive.

What Copyright Lawsuit Is Google Facing Over Gemini?

On July 10, Hachette Book Group, Cengage, Elsevier, and author Scott Turow filed a lawsuit in the U.S. District Court for the Southern District of New York, alleging that Google "brazenly copied millions of copyrighted works" to train Gemini without permission or compensation.

The publishers claim Google used their copyrighted content, available on the web and through Google Books, to train Gemini. The complaint emphasizes the unprecedented scale of the problem: "The scale and speed at which Gemini can create books and compete with human writers is unprecedented, and it can only do that because Google copied plaintiffs' and the class's works to train its AI".

This lawsuit mirrors similar copyright cases against other AI companies. In May, the same publishers teamed with McGraw-Hill and Macmillan to sue Meta over identical claims. Google has also faced previous copyright challenges; Disney issued a cease-and-desist order in December, claiming Google's Nano Banana AI image model and other video models were taking a "free ride off Disney's intellectual property" by creating AI content featuring its iconic characters.

Copyright remains one of the most contentious legal issues in generative AI. Companies building AI models require vast quantities of data to improve performance, and much of that data is copyright-protected. When AI companies scrape data from the open web or acquire it through other means, lawsuits typically follow.

However, the legal landscape has been mixed. In two major copyright lawsuits against Anthropic and Meta last year, courts sided with the AI companies. But both judges were careful to note that future cases could swing the other way. The publishers argue in their complaint that "Copyright law applies to AI companies, including Google, with the same force as every other company that has complied with these laws for decades".

Google and lawyers for the publishers did not immediately respond to requests for comment at the time of reporting.