Originally published February 2013. Updated November 2026.
Google Images has been restructured five times across thirteen years — 2013, 2018, 2020, 2023, and 2025 — with each restructuring shifting the balance of value between the user, the search engine, and the website hosting the image. The 2013 redesign that introduced inline image previews and removed the standalone hosting-page load was the first major change. The 2018 removal of the "View Image" button — the change that triggered the second wave of webmaster controversy — was the second. The 2020 introduction of structured image data, the 2023 launch of Search Generative Experience (SGE), and the 2025 integration of AI Overviews with multimodal image responses have continued the trajectory: each redesign has progressively shifted Google Images from a directory of images hosted across the web toward an AI-generated visual answer surface where the underlying hosting website becomes less relevant to the user's query.
The 2013 Redesign — Where the Argument Started
In February 2013, Google announced a redesign of Google Images that introduced several user-facing changes: larger image previews, keyboard-driven flip-through navigation, expanded metadata display (image title, source domain, image size), and a domain link that became clickable as a navigation target. Google framed the changes as user-experience improvements and stated in test data that the changes had produced a "net increase in the average click-through rate to the hosting website."
The webmaster community pushed back immediately. The redesign removed the older "load source page in background" behavior, which had inflated traffic figures for some publishers who counted background loads as page visits. Some publishers reported substantial drops in measured image-search traffic after the 2013 redesign. Others argued that the previous traffic figures had not reflected genuine user engagement and the 2013 design was structurally more honest. The substantive question — was the redesign a net win or a net loss for publishers — was difficult to answer because the publishers most affected were the ones whose previous traffic had been inflated by the technical quirks the new design eliminated.
The 2013 redesign was the moment Google Images began shifting from a "directory of images you can find on the web" toward "an image-discovery surface where the visit to the source site is increasingly optional." That shift has continued in every redesign since.
The 2018 View Image Removal
In February 2018, Google removed the "View Image" button from Google Images search results. The button had allowed users to load the full-resolution version of an image directly without visiting the source page. The removal was triggered, Google said, by a settlement with Getty Images over a years-long dispute about whether the button facilitated unauthorized image use. The settlement included Google's agreement to remove the button as part of broader changes to Google Images' handling of professional and stock photography.
The removal pushed users who wanted the full-resolution image to click into the source page — restoring some of the traffic dynamic that the 2013 redesign had reduced. But the same removal also intensified the structural pressure on publishers: with thumbnails and metadata visible in the search results page and the full image available only with a click-through, many image searchers continued their search rather than visit the source page. The 2018 change was framed as a publisher win but operated as a continuation of the trajectory toward image search results that resolved the user's need without requiring source-page traffic.
The 2020 Structured Image Data Expansion
In 2020, Google launched structured data support for images, allowing publishers to mark up images with schema.org metadata that Google would then surface in search results — author, license, creator, copyright holder, original URL, and other attribution data. The change was a publisher-friendly response to longstanding complaints about attribution and image-license violations.
The structured data layer also positioned Google Images for the AI era that was about to arrive. Schema.org markup is one of the primary signals AI engines use to understand and cite images and visual content. The 2020 change quietly built the infrastructure for the AI Overviews integration that would follow three years later.
The 2023 Search Generative Experience (SGE) Launch
In May 2023, Google launched Search Generative Experience (SGE) as a Labs-tier experiment, layering AI-generated summary answers on top of standard search results pages for select queries. The 2023 SGE included image-aware capabilities — for queries where visual information was relevant, SGE would surface AI-generated visual context alongside the conventional image search results.
For Google Images specifically, the 2023 launch meant that an image search for a recipe, a how-to procedure, a product, or an explanatory diagram could return a generated answer that incorporated visual context without requiring the user to evaluate the underlying image-hosting sites. Publishers who had spent two decades optimizing for image search ranking were now competing with a generative layer that was being trained on the same content the publishers had hosted.
The 2024 AI Overviews General Rollout
In May 2024, Google launched AI Overviews as the general-availability replacement for SGE, rolling the feature out to US users and expanding through 2024 and 2025 to additional markets. AI Overviews appear at the top of search results pages for commercial, informational, and increasingly visual queries. For image searches, AI Overviews integrated visual responses generated from the Google Images index plus the broader Google search index.
The launch generated immediate publisher concern. SEO industry analysts reported that AI Overviews appearance correlated with reduced click-through rates to underlying source sites, sometimes substantially. Google disputed the worst publisher claims and pointed to data showing that AI Overviews drove engagement to underlying sources in many cases, but the structural trajectory was clear: search results where AI Overviews appear are search results where the user's need can be resolved without source-site clicks more often than search results without AI Overviews.
The 2025 Multimodal Integration
Through 2025, Google integrated multimodal capabilities into AI Overviews — the ability to combine image, video, text, and structured data into single AI-generated responses. For Google Images specifically, the multimodal integration means an image search can return an AI-generated response that incorporates relevant images, generates synthetic explanatory imagery where useful, surfaces structured product data, embeds video clips, and includes text summary — all without requiring source-site traffic.
The 2013 redesign that started the publisher pushback was a relatively modest change in retrospect. The 2025 multimodal AI Overviews integration is a structural recomposition of what Google Images is. The product is no longer primarily a directory of images you can find on the web. It is increasingly an AI-generated visual answer surface where the source websites function as training and citation inputs more than as destinations for user traffic.
What Publishers Should Do in 2026
The 2026 publisher strategy for Google Images is functionally different from the 2013 publisher strategy. Five disciplines define the work.
Structured data first. Every image on a publisher site should ship with full schema.org markup: ImageObject, creator, copyright holder, license, original URL, caption, content location, and any other attribution fields applicable to the image type. Structured data is the primary signal AI engines use to cite images, and the publisher who marks up their content earns priority citation positioning.
Original imagery as competitive moat. The publishers who win in the AI-engine era are the ones who produce original photography, original illustration, original data visualization, and original explanatory imagery. Stock photography and licensed third-party imagery is now operationally subordinate to the AI engines' synthetic generation capabilities. Original visual assets are the asset class that cannot be synthesized away.
Citation Share measurement. Publishers should measure their Citation Share inside Google AI Overviews, ChatGPT, Claude, Perplexity, and Gemini for the visual queries that matter to their business. The measurement is the leading indicator for whether the publisher's visual content is being cited at the rate that supports its business.
Image SEO discipline continues to matter. Traditional image SEO — alt text, file names, image sitemaps, accessible captions, page-level relevance signals — continues to matter and increasingly serves as input to AI engines. The 2013 image SEO playbook has not been retired; it has been extended.
License clarity and rights management. The Getty Images dispute that triggered the 2018 View Image removal was a preview of the broader rights-management disputes that AI engine training and citation has created. Publishers should be explicit about image licensing, attribution requirements, and rights-management on every visual asset, and should consider participating in industry efforts to ensure AI engines respect publisher licensing terms.
Google Images has been substantially restructured five times: 2013 (inline previews, removal of source-page background load), 2018 (View Image button removal), 2020 (structured image data expansion), 2023 (Search Generative Experience launch), and 2024-2025 (AI Overviews general rollout and multimodal integration).
Why did Google remove the View Image button?
Google removed the View Image button in February 2018 as part of a settlement with Getty Images over a long-running dispute about whether the button facilitated unauthorized image use. The settlement included broader changes to Google Images' handling of professional photography.
What is Google AI Overviews?
AI Overviews is Google's AI-generated summary answer that appears at the top of search results pages for many queries. Launched as the general availability replacement for Search Generative Experience in May 2024, AI Overviews integrate visual, textual, and structured data into single AI-generated responses.
How does AI Overviews affect image search publishers?
AI Overviews appearance correlates with reduced click-through rates to underlying source sites, sometimes substantially. The product is shifting Google Images from a directory of images hosted across the web toward an AI-generated visual answer surface where source websites function more as citation inputs than as traffic destinations.
What is schema.org markup for images?
Schema.org provides the ImageObject type and associated properties (creator, copyright holder, license, content location, caption, etc.) that publishers can use to mark up images with structured metadata. The markup is one of the primary signals AI engines use to understand and cite images.
What should publishers do about Google Images in 2026?
Five disciplines: structured data first on every image, original imagery as competitive moat, Citation Share measurement inside AI engines, continued image SEO discipline, and clear license and rights management on every visual asset.
Reported by the Everything-PR Editorial Team.
