Lesson 11 · Activity 03

WCAG Check on AI-Generated Content

AI-generated content fails accessibility in predictable ways. In this lab you'll audit a sample AI-made module against WCAG 2.2, flag every gap with a priority, and fix the most critical failing element — the pre-publish habit that keeps your materials usable by every student.

Work through it

0 of 6 done

Review the sample module and its four elements.
Match each element to a WCAG principle (Perceivable / Operable / Understandable / Robust).
Name the gap and assign a priority for each.
Write a corrected version of the most critical failing element.
Write the reusable pre-publish checklist you’ll run every time.
Compare to the model answer.

The sample module to audit

A colleague generated this with AI and is about to publish it:

A lecture summary at a dense graduate reading level (long sentences, heavy jargon).
Three inline images (two charts, one diagram), each with alt text "image".
A data table of weekly scores with no header-row markup.
A feedback rubric using light-grey text (#BBBBBB) on white, with pass/fail shown by red/green color only.

Your audit

Element	WCAG principle	Gap & priority & fix
Lecture summary (dense reading level)
Images with alt="image"
Table with no header markup
Rubric: grey text + color-only pass/fail

Your correction of the most critical element

Your reusable pre-publish checklist — the checks you’ll run before publishing any AI-generated content

Saved automatically on this device.

Reveal model answer

Element	Principle	Gap	Priority / fix
Images alt="image"	Perceivable	Non-descriptive alt text	HIGH — write descriptive alt for each chart/diagram
Rubric grey #BBBBBB on white	Perceivable	Contrast ~1.9:1, fails 4.5:1	HIGH — darken text to meet 4.5:1
Pass/fail by color only	Perceivable / Understandable	Color is the only cue	HIGH — add a text/icon label, not just red/green
Table without header markup	Robust	Screen readers can't associate headers	MEDIUM — mark the header row (th/scope)

Most critical fix: the rubric uses color alone and fails contrast. Darken the text to a passing color and add an explicit "Pass" / "Needs work" label beside each color cue, so the meaning survives for colorblind users and screen readers. (The summary's reading level is also a real Understandable issue — ask the AI to simplify.)

Your reusable pre-publish checklist (copy and keep):

Before publishing any AI-generated content to students:
[ ] Perceivable — descriptive alt text; text contrast at least 4.5:1 (3:1 for large/bold).
[ ] Operable — works with a keyboard, not just a mouse.
[ ] Understandable — logical heading order; reading level fits the students.
[ ] Robust — real headings and table headers so screen readers can parse it.

Takeaway: Most failures came straight from the AI's defaults — generic alt text, low contrast, color-only cues. That's exactly why a quick four-principle pass before publishing is non-negotiable for AI-generated content.