The Accessibility Automation Trap

In the rush to become compliant with upcoming accessibility standards, many organizations are likely to turn to “quick fix solutions” like automated scanners and plugins. In a world so heavily reliant on artificial intelligence, this is expected—particularly from organizations with little resources, funding, or runway to navigate new accessibility guidelines.

I, myself, use many tools to help me in my day-to-day work. And though many of these tools are incredibly helpful, relying on them to accomplish accessibility and remediation work is akin to asking a spell-check to write a great novel; it may catch your grammatical errors, but it will completely miss the plot.

The truth is, accessibility can’t be reduced to a checklist of technical requirements because it is a human experience.

Let’s explore why technology along will always fall short of the goal without a knowledgeable human at the helm.

#1. Context is King

Automated tools and technology are excellent at detecting objective errors, such as a missing image tag or meta description, however they struggle with the subjective aspect of remediation.

One example of what this looks like in practice is regarding alternative text (“alt” text for short). An automated tool can absolutely flag that an image is missing alternative text. It can even write up some alternative text for you… But it lacks the ability to tell you whether that alternative text is helpful within the context of the image and its purpose.

This this photo, for example:

An automated tool may describe this photo as “An older woman stands in front of a building and smiles”, while a human who understands the context behind the photo would say “Co-Founder of Woozles bookstore, Liz Crocker, stands in front of the original building, opened in 1978.”

Another example of how this looks in practice is the meaning of content structure. Automated tools can verify that headings exist in your report or document, but they cannot determine whether the logical flow of the document makes sense when using a screen reader. This is something a human must double check.

#2. Accessibility is More Than Technical… It’s Functional

Artificial intelligence lacks the tacit knowledge and considerations required for accessibility improvements to be meaningful. Take the following accessibility features, drawn from WCAG (Web Content Accessibility Guidelines) for example:

Colour Contrast

When it comes to colour contrast guidelines, automated tools can measure hex code ratios and tell you whether a colour combination is technically compliant, however most will fail to tell you whether that same combination is accessible for individuals with a variety of visual impairments, and a human can identify whether a vibrant colour combination may cause migraines or sensory overload for your clients or customers.

Keyboard Navigation

An automated tool can tell you is a button is focusable, however a professional can feel if the tab order is frustrating or illogical.

Forms

An automated tool can check for labels, and perhaps even assign ones for you. A professional, however, can ensure error messages actually help a user fix the problem so that they can complete and submit the form.

#3. The Human Experience Is Not Technical

There are many situations in which your website, report, or content is “technically” compliant, and passes automated tests, but is practically unusable.

Real customers, clients, and users don’t interact with your code; they interact with screen readers, switch controls, voice-to-text software, and other features.

While automated tools can simulate these interactions, they cannot anticipate, replicate, nor understand the frustration of a keyboard trap or the confusion of a poorly labelled document or interactive database.

Relying exclusively on technology treats accessibility as a list to check off, or a thing to “fix” rather than a living, breathing person to be serviced.

Your goal as a brand or organization should not be technical accessibility, but rather functional accessibility—leveraging real human experiences and knowledge.

Technology can absolutely be the tool to help you identify obvious hurdles or missing tags, but humans should be the architects of your accessibility strategy, remediation, and education.