The 15-second version
- What's happening: Google is putting AI image generation (Nano Banana model) directly into AI Overviews, announced July 14, 2026 for Google Images' 25th anniversary.
- Rollout: over the coming weeks, in English, across regions where image creation is already live in AI Mode.
- Alongside it: a redesigned Google Images homepage, an immersive, personalized, real-time gallery (desktop U.S. first).
- The SEO stake: one more zero-click surface, and head-on competition for image search.
On July 14, 2026, the 25th anniversary of Google Images, Google announced it is bringing AI image generation directly into Search AI Overviews, according to an announcement on its official blog. Its in-house Nano Banana model turns a text prompt into a custom visual, built "completely from scratch," without the user ever leaving the results page.
In Google's words, the feature "transforms a simple text prompt into a high-quality, custom visual made completely from scratch." The rollout starts over the coming weeks, in English, for regions where image creation is already available in AI Mode. In parallel, Google unveiled a full redesign of the Google Images homepage: a "dynamic, immersive" gallery of images from across the web, updated in real time and "intelligently tailored" to each signed-in user's interests.
Why it matters
Until now, image search was one of the few surfaces where Google necessarily sent users to third-party sites: to see and grab a visual, you had to click. By generating the image on demand inside the AI Overview, Google removes that detour. As Search Engine Journal notes, "AI Overviews already answer many queries on the page without a click", adding generated images extends that self-contained logic even further.
The backdrop is already tense. We recently covered how 68% of searches now end without a click, and a study analyzed in our piece on the 42% drop in organic traffic tied to AI Overviews shows how the direct-answer surface eats into editorial traffic. Image generation opens the same front, this time on visual traffic.
The detail that counts: when Google introduced visuals into AI Mode, its systems "did not explicitly distinguish real photos from AI-generated images," a company spokesperson acknowledged to Search Engine Journal. In other words, a fabricated image and a real photo can sit side by side in the same answer, a trust problem for brands.
What it changes for small and mid-sized businesses
Three concrete consequences, even though the rollout starts in English and will reach other markets later:
- "Image search" traffic will contract. Sites that captured visits via Google Images (e-commerce, real estate, tourism, visual crafts) should expect erosion, especially on generic queries where a generated image is enough for the user.
- Provenance becomes strategic. If the AI generates a "product" or "place" visual instead of your real photo, your visual differentiation fades. Authenticity signals (C2PA marking, metadata, original dated photos) gain value, a topic we detailed on AI image watermarking.
- Visual visibility runs through citation, not clicks. As with text, the goal becomes being the source the AI shows and attributes, rather than the link you hope gets clicked.
Selling products, places or visual creations? The free Cicéro diagnosis measures your exposure to this shift, and pinpoints the visuals and pages to protect first.
What to do now
Without waiting for the feature to land in your market:
- Audit your reliance on Google Images. In Search Console, isolate the share of traffic coming from the Images tab. That's your direct exposure.
- Strengthen the uniqueness of your visuals. Original, contextual photos, with captions and precise descriptive
alttext, what AI can't fabricate without losing accuracy. - Structure images for citation.
ImageObjectmarkup, clear licensing, attribution, and anchoring to strong text content. This is Generative Engine Optimization applied to visuals. - Document authenticity. Metadata, capture date, provenance marking: proof that distinguishes your real photo from a generative render.
Our take
Google is closing the last exit door of zero-click: after text answers, now images. For an SMB, the conclusion is the same as for text, betting on generic is a losing game. What stays defensible is the authentic visual: dated, attributed, backed by real expertise. AI can fabricate a "good enough" image; it can't fabricate your proof.
What this article doesn't cover
The exact rollout timing for non-English markets isn't disclosed yet, and Google has published no figures on the feature's impact on publisher traffic, it's too early. We don't address the legal angles here (copyright of generated images, liability for a misleading visual), which call for specialist advice. The recommendations above are strategic directions, to be adapted to your sector.
Frequently asked questions
What did Google announce on July 14, 2026?
Is it available everywhere?
What is the SEO impact?
Sources
- → The Keyword (Google), official announcement of Google Images' 25th anniversary and image generation in AI Overviews (July 14, 2026)
- → Search Engine Land, "Google AI Overviews now lets you create images" (July 14, 2026)
- → Search Engine Journal, "Google Revamps Image Search, Adds Generation To AI Overviews" (July 14, 2026)
Growth and SEO content strategist, I founded Cicéro to help businesses build lasting organic visibility, on Google and in AI-generated answers alike. Every piece of content we produce is designed to convert, not just to exist.
LinkedIn