Alt Text Is About Meaning — Not Pixels (Extended)

One of the biggest misunderstandings in accessibility is that alt text should simply describe what you see. In practice, literal descriptions often fail users because they preserve pixels, not purpose.

Story: when “correct” still misleads

I once worked with a team who followed the rule they’d heard everywhere: “alt text should be a short description of the image.” They shipped dozens of literal labels such as “Image of a child,” “Image of a building,” and—my favourite—“Image of orphan and Batman.”

The picture showed a crying orphan next to a brick wall; his shadow on the wall was Batman. The intended message was: “The difference between ordinary and extraordinary is just a little extra.” A literal label (“image of orphan and batman”) was technically descriptive but functionally misleading—it hid the idea the image was communicating.

Missing vs. wrong alt: the harm isn’t equal

• Missing alt on an informative image is an exclusion: screen reader users hear nothing useful and may have to guess. That’s bad.
• Wrong (or literal) alt can be worse: users are misled into believing an untrue or incomplete meaning. That’s not just unhelpful—it’s a form of misinformation. In critical contexts (health, finance, safety), misleading text can increase risk.

Alt text should communicate intent: why the image is present and what the user is meant to take away. Ask: “What idea is this image trying to convey?”

Practical guidance for teams

• Decorative imagery: use empty alt (alt="") so screen readers skip it.
• Informative imagery: write the message, not the pixels (what a sighted user would grasp at a glance).
• Functional imagery (icons/buttons): describe the action (e.g., “Search”, “Download the report”).
• Complex graphics: pair a succinct alt with a nearby or linked longer description.
• Don’t prepend “image of …” or “picture of …”. Assistive tech already announces role; redundancy adds noise.

Fact box: Captions (A) vs. Sign language (AAA)

Teams often ask why captions are a baseline requirement while sign language is not. In the Web Content Accessibility Guidelines (WCAG):

• Captions (Prerecorded) — Success Criterion 1.2.2 — Level A: time-synchronised text alternatives for spoken dialogue and meaningful audio are required for prerecorded media.
• Sign Language (Prerecorded) — Success Criterion 1.2.6 — Level AAA: providing a sign-language translation of the audio track is an enhanced requirement.

Why the difference? Text captions ensure core information is available to a broad range of users (Deaf, hard of hearing, those in noisy or quiet environments, second-language readers), and they’re machine-indexable and searchable. Sign language can convey prosody, emphasis and cultural nuance that text may miss, offering richer comprehension for native signers—but it’s considered an advanced enhancement, so it sits at AAA. Where your audience includes a high proportion of signers, adding sign-language tracks can substantially improve understanding.

Checklist: writing meaningful alt text

1. Identify the purpose of the image in its context.
2. Write what the user needs to know, not what you can see.
3. Avoid boilerplate openings (“image of …”).
4. Keep it concise; link to more detail when necessary.
5. Test with a screen reader—or at least read your alt text aloud in place of the image.

Bottom line: literal alt text checks a box; meaningful alt text preserves the experience.