Alt text: the line Google drew in 2022, that 53% of sites still skip
What Google actually drew a line around in 2022
For years, "does alt text help SEO" got a vague answer, because the honest answer splits into two systems that most advice collapses into one. Google Search Advocate John Mueller addressed it directly during a Search Central SEO office-hours hangout recorded on 18 March 2022, as reported by Search Engine Journal: "It's not that the webpage in the text web search would rank better because it has an image." Alt text, in other words, does not act as a hidden keyword booster for your ordinary Google ranking. A reader pushed back on Twitter, and on 26 March 2022 Mueller clarified the point rather than walked it back: "Google sees alt-text as a part of the page (you can search for it and find it in the snippet)." Alt text is indexed as text on the page, which is a real, if modest, effect. It is just not the ranking lever a decade of SEO checklists implied it was.
Image Search is the part that is a genuine ranking factor
The system where alt text is not a side effect but the main input is Google Images. Google's own image SEO documentation, at developers.google.com/search/docs/appearance/google-images, states plainly: "The most important attribute when it comes to providing more metadata for an image is the alt text." The same page explains why: "Google uses alt text along with computer vision algorithms and the contents of the page to understand the subject matter of the image." That is a direct description of alt text doing ranking work, just inside Image Search rather than the main web results. The documentation adds two practical details worth keeping. If an image doubles as a link, its alt text functions as the anchor text for that link, which is one of the few places alt text quietly touches ordinary SEO after all. And stuffing it with keywords backfires: Google's guidance says keyword-stuffed alt text "results in a negative user experience and may cause your site to be seen as spam," which turns a free accessibility win into a liability for nothing in return.
The compliance gap: the same failure, seven years running
None of this is new information. Google has been explaining the image-search-versus-web-search split for years, and the underlying HTML attribute has existed since the 1990s. What has not changed is how many sites bother. WebAIM, the accessibility research group behind the annual Million report, tested the top 1,000,000 home pages with data pulled in February 2026, and found missing alternative text among the most common accessibility failures on 53.1% of them, down only slightly from 55.5% in the 2025 report. That 2025 report breaks the same problem down a second way: 18.5% of every image found across those million home pages, not just the pages but the individual image tags, had no alt text at all. However it is sliced, the pattern holds. This single failure type has sat among WebAIM's six most common accessibility problems for seven consecutive years, through alt text being explained, re-explained, and built into every CMS's media library as an editable field. The fix costs nothing and takes a sentence. Half the web still skips it.
Where this intersects with AI search, stated honestly
The newer question is whether alt text matters to the AI assistants and answer engines that increasingly sit between a page and the person reading it. Here the evidence runs thin fast. Some vendor content asserts that alt text is critical for how AI systems "interpret and understand" a page, including a DubBot piece on the subject published 19 February 2026, but claims like that one cite no data, no methodology and no named study, only reasoning by analogy to how screen readers and NLP systems work. We could not locate a single public, dated, methodologically disclosed study that measures whether ChatGPT, Gemini, Perplexity or Google's AI Overviews weight alt text specifically when deciding what to cite about an image. That gap is worth stating plainly rather than papering over with the WebAIM or Mueller figures above, which describe web search and accessibility, not AI citation behavior. What is defensible without a study: the same logic that governs where AI engines actually quote from on a page applies here too, because a model can only cite what it can read, and for a crawler or model that does not render or interpret the image itself, alt text remains the only description of it that exists at all.
What alt text still buys you in 2026
Strip away the inflated SEO-ranking claims and what is left is still worth doing, for reasons that do not need a disputed study behind them. It is the only description a screen reader has to work with, which is the use case Google's own guidance leads with and the one most checklists bury at the bottom. It remains a direct, named ranking input for Google Images, a real source of traffic on plenty of sites that sell anything visual. It doubles as anchor text whenever an image is wrapped in a link, a detail most sites never set deliberately. And unlike almost everything else in this field, it is free: one honest sentence per image, written to describe what is actually there rather than stuffed with keywords. Oversized, unoptimized image files cause a separate and more consequential problem, since they are the most common reason pages fail Core Web Vitals, a topic our piece on whether Core Web Vitals still move AI citations covers on its own; alt text does not touch file size at all, so fixing one does nothing for the other, and both are worth doing regardless.
What we could not verify
No public source we could find discloses how, or whether, any major AI engine weights alt text when selecting or describing images in an answer, so the AI-citation framing above is stated as an open gap, not a measured effect. We have WebAIM's 2026 figure for the share of homepages with a missing-alt-text failure, 53.1%, but not yet an updated 2026 figure for the separate per-image statistic, so the 18.5%-of-all-images figure cited above is specifically the 2025 report's number, not this year's. And Google has never published a numeric weighting for how much alt text affects an Image Search ranking the way it once did, briefly, for HTTPS; "the most important attribute" is a qualitative claim in Google's documentation, not a quantified one.
A check worth running this week
- Pull up your 10 highest-traffic pages and inspect the image tags. Look for alt="" that is empty by accident rather than by design, and alt text that is just the filename.
- Confirm decorative images use an intentionally empty alt="", not a vague filler word. A spacer or background flourish described as "image" or "icon" helps nobody and clutters a screen reader's output.
- Run your homepage through WebAIM's free WAVE tool. It will tell you, in under a minute, whether you sit above or below this year's 53.1% average, and exactly which images are missing text.
Sources
Google's position on alt text and the image-search-versus-web-search split is documented in Google's own Google Images SEO best practices and reported from John Mueller's 18 March 2022 Search Central office hours and his 26 March 2022 Twitter clarification in Google: Alt Text Only A Factor For Image Search, Search Engine Journal, 22 March 2022. The accessibility figures come from WebAIM, The WebAIM Million 2026, data pulled February 2026, page last updated 30 March 2026, and the prior-year comparison and per-image figure from The WebAIM Million 2025, data pulled February 2025.
Common questions
Is alt text a Google ranking factor?
For Image Search, yes, directly. Google's documentation calls it the most important attribute for helping its systems understand an image. For ordinary web search, Mueller said in 2022 that a page does not rank better for having an image with alt text, though alt text is still indexed as part of the page's text.
Do AI engines use alt text to understand or cite images?
No public, dated study confirms this one way or the other. Multimodal models can look at the image itself, which may reduce reliance on alt text, but no engine discloses its weighting. Alt text remains the only text description available to any crawler or reader that cannot view the image directly.
What percentage of websites are missing alt text in 2026?
53.1% of the top 1,000,000 home pages, per WebAIM's 2026 Million report, down only slightly from 55.5% in 2025. It has been among WebAIM's six most common accessibility failures for seven consecutive years.
Structured data runs on the same logic as alt text: a plain, honest statement of what is actually on the page, in a form a machine does not have to guess at. See our guide to schema markup that earns AI citations for the rest of that practice.
Fix the easy wins across every page
On-Page Optimization rewrites titles, metas, headings and image attributes page by page, so the free fixes like this one stop sitting half-done across your site.