Smart companies are always looking for ways to leverage automation to keep their businesses efficient and competitive. However, while we should always be looking to push the boundaries of what is possible, there are some things which simply should not or cannot be entirely automated. At present, one of these things is website accessibility.
Just like you would not use a machine to tell you whether or not your website is designed well, or whether it provides a pleasant experience for your users, you should not rely on a software program, no matter how smart it may claim to be, to verify whether your site is accessible. The reasoning behind this is simple: ultimately it is a human who will have to experience your website, and until we are able to teach computers what constitutes a good experience for a human, we cannot rely on them to make that determination.
Whereas a disabled user is likely to access your site in a unique manner—such as with a screen reader, without using a mouse, or with a sip-and-puff switch—a scanning tool can only attempt to simulate these conditions.
Of course, this is not to say that automated accessibility testing tools are without their uses. They can be helpful in identifying simple issues. For example, every form field should have a label associated with it, and a computer program can very easily determine whether or not this is the case. However, not only does there need to be a label associated with each form field, but it also needs to contain some content which explains what the field does. This last is relatively straightforward for a human to establish, but not so much for a machine.
Another example is alt attributes. An alt attribute is how a web developer attaches a description to an image in HTML. An automated accessibility scanning tool can very easily determine whether or not an image has an alt attribute, but what it can’t do is ensure that the description matches the image, and that it does not leave out any pertinent details.
In short, automated accessibility tools can help developers identify some of the low hanging fruit of accessibility issues on their sites, but they certainly won’t identify all the issues, and the output needs to be verified by someone with a good understanding of accessibility to weed out any false positives, and find the remaining false negatives. Failing to properly do so inevitably leads to wasted effort, and false confidence in the accessibility of the site.
We’re not the only ones making this claim. Here’s what the U. S. General Services Administration has to say about this:
Automated scanning provides an important first-pass ‘screening’ that can identify if a website is not accessible or does not comply with accessibility standards by testing for the absence of valid required elements and/or attributes.
Automated scanning cannot determine if a website is accessible or conformant with accessibility standards. Many accessibility checks require human judgment and must be evaluated manually using different techniques.
While automated scanning tools can be a good place to start for somebody who has a solid background in accessibility, they should never be your only approach to ensuring compliance with the ADA, WCAG 2.0, Section 508, or any other technical or legal accessibility standard.