We covered difficult accessibility in my courses a couple of weeks ago. We talked about digital elements that pose accessibility challenges like maps, website slideshows, digital signage, etc. We were then asked if we thought these things would always be challenging, and I said no. While all the examples will always be difficult to make accessible, the rapid advancement of AI is going to be key in bridging the gap between these visual mediums and the ability of nonvisually oriented people to interact with them. Even before the advent of large lanuage models and their availability to the masses, I already had tools that allowed me to switch the inputs on my TV, find an unmarked button on my microwave, etc., but they took too long to be useful. There’s also a tool for iPhone, screen recognition, that often allows me to interact with apps that would otherwise be inaccessible. The feature doesn’t fix the app, but it usually lets me work through steps in the setup or signup process that would otherwise have locked me out entirely. Generative AI has its problems, but the increase in ease of accessibility is definitely an up.