Why digital Accessibility will fail without Automation
Automatic accessibility has sometimes rightly fallen into disrepute. This is mainly due to the so-called accessibility overlays, which usually do not contribute to the accessibility of a website, but often even worsen it.
However, my impression is that many persons are structurally conservative or fear that they could lose orders through automation. You can see that clearly in the translation industry. Tools like DeepL or Google Translate have gotten better and better in recent years for texts in everyday language. Specialist texts are a different topic.
Yes, there are crappy translations that are worse than the literal translations of late 90's tools. As a rule, however, it is about useful text, not about literary quality.
And we need good translation tools. There are persons who don't understand the original language well enough. At the same time, however, there is a need for many persons to understand texts in other languages.
We find similar problems in accessibility. Over the next few years, the need for testing will increase dramatically as some organizations take accessibility more and more seriously. In my opinion, the test routine that has been established today with the inflated BITV tests is not effective. The thousands of man-days that are put into this can also be used more sensibly. The current process is this: we do it wrong, have it tested, do it a little less wrong, have it tested and so on.
Without automation, we will hardly achieve any of the accessibility goals. Developers and designers need tools that are built directly into their workflows and help them avoid, identify and fix errors. We need speech-to-text engines that automatically subtitle films or convert podcasts to text. We need tools that also enable laypersons to write more understandable texts.
With today's resources, it's not possible to create as many accessibility professionals as we need. There is a shortage of young talent - how many professionals under 35 do you know - and the non-accessibility professionals are already busy trying to keep up with their own discipline.
What we are currently feeling is not so much the lack of accessibility experts. We have an overall lack of human working power and financial resources. It's absurd to think that there are still persons doing things manually that could be completely automated.
Now no one is claiming that automation will solve all problems. We definitely need better tools. However, I have long suggested using the three methods of automation, heuristic expert analysis, and user feedback in combination.
Technology will not solve all problems
At the same time, be warned against naive technological optimism. There are two types of predictions: realistic and utopian. In principle, realists simply update current developments: computers are becoming a little faster, cars a little more economical, and so on. Utopians try to foresee what does not yet exist. Often they follow a naïve forced optimism, recognizable in the current environmental policy. If we replace all combustion cars with electric cars, do away with fossil fuels for heating and only heat with electricity, we will need enormous amounts of electricity and storage capacities. These cannot be achieved with the current technical innovation.
Neither will we develop technical solutions that solve all of today's accessibility problems. At least not in the lifetime of today's adults.
- Artificial Intelligence and digital Accessibility
- ChatGPT and digital Accessibility
- Cognitive Bias and Accessibility
- Can Software reduce discrimination
- Why Technology will not solve all accessibility problems
- Why digital Accessibility should be based on Open Source
- Digitization and accessibility - opportunities for disabled persons