{"id":10305,"date":"2026-03-25T11:17:00","date_gmt":"2026-03-25T15:17:00","guid":{"rendered":"https:\/\/blogs.swarthmore.edu\/its\/?p=10305"},"modified":"2026-03-25T08:15:06","modified_gmt":"2026-03-25T12:15:06","slug":"unpacking-the-accessibility-trap-of-google-slides-beautify-feature","status":"publish","type":"post","link":"https:\/\/blogs.swarthmore.edu\/its\/2026\/03\/25\/unpacking-the-accessibility-trap-of-google-slides-beautify-feature\/","title":{"rendered":"Unpacking the accessibility trap of Google Slides\u2019 \u201cBeautify\u201d feature"},"content":{"rendered":"\n

If you use the Google Suite, you\u2019re no doubt familiar with the daily creep of Gemini and other Google-based AI tools into your documents. In Google Slides, you\u2019re constantly prompted to \u201cbeautify\u201d your slide using Nano Banana, Google Gemini’s image-generation tool (indicated by the small banana logo). <\/p>\n\n\n\n

\"A<\/figure>\n\n\n\n

While building a slide deck recently\u2014on the effects of AI overuse, funnily enough\u2014I decided to test it out on the slide below, using the default output with no prompt injection. The results are alarming for many reasons, but most importantly because the \u201cbeautify\u201d process strips all of the accessibility features from the slide.<\/p>\n\n\n\n

\"A
My original slide.<\/figcaption><\/figure>\n\n\n\n

It turns out \u201cbeautify\u201d is a bit of a misnomer. It appears to feed the contents of your slide into an image-generation tool with the prompt \u201cBeautify this slide.\u201d It does not generate text boxes or shapes. The final \u201cbeautified\u201d slide is a slide-sized image\u2014not an editable slide<\/em>.<\/p>\n\n\n\n

Not only that, the feature rephrased my text, fictionalized brain scan data, and contained spelling errors. The randomly generated brain scans were factually inaccurate and certainly misleading. Furthermore, since the new \u201cslide\u201d was actually an image, I couldn\u2019t alter the text if I wanted to. I couldn\u2019t fix the spelling errors or remove the inaccurate graphic either. <\/p>\n\n\n\n

\"An
The slide after “beautification” with the default prompt. Notice the scale in the upper right is labeled \u201csignificaty\u201d instead of \u201csignificantly.\u201d The brain scans also do not seem to represent the data in the original graphic at all.<\/figcaption><\/figure>\n\n\n\n

Making the content not editable seemed to destroy the very idea of a slide as a whole. For a bit more control over the slide\u2019s contents, I tried beautifying again with the following prompt:<\/p>\n\n\n\n

Summarize the key takeaways from text on this slide into three bullet points using scholarly language. Add a border to the image but do not alter its contents. Reformat the slide's design to be easily legible and have a classy, educational design<\/code><\/pre>\n\n\n\n
That got me this image:<\/p>\n\n\n\n
\"An
The slide beautified with a more specific, engineered prompt. Note that it’s possibly less digestible than the original.<\/figcaption><\/figure>\n\n\n\n
It still suffered from the same drop in quality, most notably in the brain scan image. It originally appeared to leave the image unaltered as I requested, but the drop in quality does make some text functionally illegible. When zooming in to 200% and 400%, I found that much of the image\u2019s text had been garbled into incomprehensible lines as well\u2014so there must have been some alteration.<\/p>\n\n\n\n
Some of the \u201cscholarly language\u201d was awkward, such as \u201cThe experimental group consistently demonstrated underperformance\u201d instead of \u201cThe experimental group underperformed.\u201d The third bullet point particularly was flowery to the point of hurting understanding: cognitive decline \u201cculminating in increased reliance on verbatim transcription\u201d was way more wordy (and hard to follow) than the original \u201cChatGPT users [were] often resorting to copy-and-paste.\u201d Plus the dense lines of text didn’t seem to help much with the “legible” trait I requested.<\/p>\n\n\n\n
Even beyond the language itself being inaccessible to read, this of course renders most screen readers useless <\/strong>as there is no longer text for screen readers to ingest. It also eliminates formatting such as titles, subtitles, and paragraph labels that screen readers and keyboard users rely on to navigate visual material. Despite generating an image, it does not generate accompanying alt text. After beautifying, the alt text was merely populated with the prompt (for the first version, the description was simply \u201cBeautify this slide\u201d). Contrast was low in parts of the first slide and text was extremely dense on the second slide. <\/p>\n\n\n\n
When making any slides, we recommend Grackle<\/a>, an accessibility checker extension that checks for proper structure and image descriptions for screen readers. (Although it should be noted Grackle won\u2019t catch the inaccessibility of beautified slides, because its screen-reading and slide-reading capabilities don\u2019t apply to the image.) If your Google Slides aren\u2019t feeling beautiful, you can try using a beautified slide as inspiration (although you must check the text and images for the insertion of incorrect, misleading, or misinterpreted material). Or you can do what I do: build a deck using one of Google\u2019s premade templates. Always evaluate the accessibility of your tools!<\/p>\n","protected":false},"excerpt":{"rendered":"
If you use the Google Suite, you\u2019re no doubt familiar with the daily creep of Gemini and other Google-based AI tools into your documents. In Google Slides, you\u2019re constantly prompted to \u201cbeautify\u201d your slide using Nano Banana, Google Gemini’s image-generation … Continue reading Unpacking the accessibility trap of Google Slides\u2019 \u201cBeautify\u201d feature<\/span><\/a><\/p>\n","protected":false},"author":88,"featured_media":10311,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_monsterinsights_skip_tracking":false,"_monsterinsights_sitenote_active":false,"_monsterinsights_sitenote_note":"","_monsterinsights_sitenote_category":0,"_jetpack_newsletter_access":"","_jetpack_dont_email_post_to_subs":false,"_jetpack_newsletter_tier_id":0,"_jetpack_memberships_contains_paywalled_content":false,"_jetpack_memberships_contains_paid_content":false,"footnotes":""},"categories":[2,71,436,186],"tags":[216,90],"class_list":{"0":"post-10305","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"has-post-thumbnail","6":"hentry","7":"category-academic-technology","8":"category-accessibility","9":"category-artificial-intelligence-2","10":"category-design-thinking","11":"tag-accessibility","12":"tag-featured","14":"fallback-thumbnail"},"jetpack_featured_media_url":"https:\/\/blogs.swarthmore.edu\/its\/wp-content\/uploads\/2026\/03\/originalslide.png","jetpack_sharing_enabled":true,"jetpack_shortlink":"https:\/\/wp.me\/ph2nPL-2Gd","_links":{"self":[{"href":"https:\/\/blogs.swarthmore.edu\/its\/wp-json\/wp\/v2\/posts\/10305","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/blogs.swarthmore.edu\/its\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/blogs.swarthmore.edu\/its\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/blogs.swarthmore.edu\/its\/wp-json\/wp\/v2\/users\/88"}],"replies":[{"embeddable":true,"href":"https:\/\/blogs.swarthmore.edu\/its\/wp-json\/wp\/v2\/comments?post=10305"}],"version-history":[{"count":9,"href":"https:\/\/blogs.swarthmore.edu\/its\/wp-json\/wp\/v2\/posts\/10305\/revisions"}],"predecessor-version":[{"id":10330,"href":"https:\/\/blogs.swarthmore.edu\/its\/wp-json\/wp\/v2\/posts\/10305\/revisions\/10330"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/blogs.swarthmore.edu\/its\/wp-json\/wp\/v2\/media\/10311"}],"wp:attachment":[{"href":"https:\/\/blogs.swarthmore.edu\/its\/wp-json\/wp\/v2\/media?parent=10305"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/blogs.swarthmore.edu\/its\/wp-json\/wp\/v2\/categories?post=10305"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/blogs.swarthmore.edu\/its\/wp-json\/wp\/v2\/tags?post=10305"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}