The Gold Standard for Digital Accessibility — And How We Get There

When I talk to people about the importance of digital accessibility, I usually mention that just 3% of the internet is accessible to people with disabilities.

It’s a sobering statistic, made worse by the fact that this number hasn’t improved in recent years — even though more organizations are aware of digital accessibility. 

The lack of progress signals that many organizations are taking the wrong approach to accessibility. Instead of leveraging the strengths of both automation and human experts, they choose one of the other — and struggle to deliver accessible experiences to every user.

Finding the Right Balance of Experts + Automation

One of the biggest challenges in digital accessibility is the size and dynamic nature of the internet.

Most websites change all the time. And each update — a new image or a pop-up banner — is a chance to introduce new accessibility issues accidentally.

Automation can identify and remediate many of these issues, but it isn’t a silver bullet. Today’s automated tools can only test for about 1/3 of the latest Web Content Accessibility Guidelines (WCAG) — the accessibility standards cited by laws like the Americans with Disabilities Act (ADA).

Unfortunately, many overlay companies give customers a false sense of security by promising to make their websites fully accessible and compliant with one line of code.

At AudioEye, we believe that automation is critical to any comprehensive approach to digital accessibility. But you still need human experts to test the remaining WCAG criteria — ideally starting with your highest-trafficked pages and user flows. 

When Paired Together, Experts And Automation Can Reduce Risk

Earlier this year, we analyzed over 900 legal claims in the accessibility industry to determine how many of them were valid — meaning the accessibility issues cited in the lawsuit could be reproduced and confirmed. 

After reviewing the lawsuits, we found that the highest percentage of valid claims (~70%) were against companies with no accessibility solution in place.

By comparison, companies that used consulting approaches and developer tools (meaning they manually fixed accessibility issues at the code level) had a valid claim rate of 50%.

At AudioEye, our unique combination of automated and custom fixes — including at-source fixes — resulted in a valid claim rate 62% lower than the nearest comparison of consulting approaches.

The Gold Standard Approach to Accessibility

So, how do we reach a point where the rate of valid accessibility claims is as close to 0% as possible?

At AudioEye, we believe it requires a multi-pronged approach to accessibility that includes more education, awareness, and testing.

First, you should take a preventative approach to accessibility. In design thinking, it’s common to consider customer needs when building a new experience. We need to get more designers, developers, and engineers thinking about accessibility from the beginning through increased awareness, understanding, and empathy.

Second, you should take a preemptive approach that prioritizes accessibility at every stage of the design and development process — and gives people the tools to test content before publishing it. An automated accessibility solution can be an essential last line of defense, but you don’t want to rely on it to fix everything. Because, as we know, it can’t.

Finally, you should take a proactive approach that remediates accessibility issues as they arise — often from content updates, personalization or targeting efforts, or design changes to your website or digital experiences. A comprehensive approach includes automated remediations, source fixes, and expert audits.

The Future of Accessibility Will Lean More on AI-Based Automation

Recently, an AudioEye study found that AI could reduce the time it took to resolve a moderately complex issue (the accuracy of link descriptions for non-sighted users) by up to 10x, with no discernable loss of quality of accuracy.

As generative AI and Large Language Models (“LLM”) continue to improve, we can use automation to identify and remediate more accessibility issues — reducing the need for people to audit so many WCAG criteria. We are leading efforts by investing heavily in AI and improved automation technology to ensure we achieve our mission of making the internet inclusive for all. 

I believe we’ll eventually reach a point where AI-based solutions can identify 70 to 80% of accessibility issues — at a degree of accuracy and authenticity indistinguishable from human experts.

Scaling Digital Accessibility Requires Admitting Our Limitations (Today)

I’m excited that more organizations are looking into the accessibility of their websites and mobile apps. But there’s still work to be done. And it starts with changing how organizations think about accessibility.

For 30 years, people have been trying to solve digital accessibility the same way, and the accessibility gap isn’t getting any smaller. We can’t continue to expect that we’ll get different results by only focusing on at-source fixes or only technology.

Digital accessibility is never going to be fully automated. There will always be a need for human experts (including members of the disability community) to share their perspectives and steer automation in the right direction.

Over time, we can get AI-based automation to a point where automated fixes are indistinguishable from ones delivered by human experts. In that case, we’ll further scale digital accessibility and make steady progress toward a more accessible, inclusive future.