How Will Machine Learning Affect Digital Accessibility?

Illustration of person reading a book and using a tablet with snake around neck

Machine Learning will be incredibly important in the advancement of digital accessibility.  But what will that advancement look it?  And what can Machine Learning actually do for web and content accessibility? 

Better, More Comprehensive Testing

Automated testing is a powerful tool for web accessibility and many great tests already exist. WAVE is a good one for websites and ACE by Daisy is a good one for EPUBs. But these tests aren’t without their limitations. And those limitations are quite significant at the time of writing.   

These tests mostly work by looking at the code that makes up the content of a website to make sure certain things exist within it. But these tests aren’t able to look at or measure the actual experience of a person with a disability using the site. For example, just because an alt-text is present on the site, doesn’t mean that the content of that alt-text is descriptive or accurate enough for a person using the site to properly interact with it. A test that utilizes machine learning could do more to ascertain the meaning and quality of the present alt-text and give a more useful diagnoses. 

Inclusive Design for Everyone 

A lot of the work of digital accessibility goes on behind the scenes.  Properly ordered semantic HTML elements for sensible underlying content structure, for example. Right now, this is work that requires someone with a technical background.  But someday, tools that utilize machine learning algorithms could be used to implement these things automatically.  A tool like this would ensure that content structure would always be properly coded even as the content on the page or document changes.  That’s because the tool would be changing the underlying code based on structural context ascertained by the machine learning algorithm.  A tool like this would allow anyone to have an accessible site. It would even work for websites built using templates or software.  

In the future, Machine Learning will help set a much higher standard for content and web accessibility.  Tools created using Machine Learning will be instrumental in making the tasks of testing for accessibility and implementing accessibility fixes less technical and less time consuming.  These improvements will surely help pave the way for greater standards of inclusiveness and accessible design. 

Scroll to top