AI Breaking Barriers

Recently, the topic of Artificial Intelligence (AI) has been a huge thing in the public consciousness; people have been talking about this technology’s environmental impact, possible benefits for humanity, and more. However, we’re overlooking one of the biggest impacts it has: helping people with disabilities participate in an able-bodied world. I’d dare to say that it’s almost as big as curb cuts—those dips in the sidewalk that were “rolled out” (pun intended) for wheelchair users, but are also used by parents with strollers, people with suitcases, etc.. AI is something of a “cyber curb cut”, by which I mean that it benefits everyone, not just those with disabilities.

Here are some examples:

Grammar and Spellcheck

We’ve all been there, right? Working on a document and doubting if the phrasing is an appropriate match for the tone we’re trying to have? I think most of us assume this kind of tool, like Grammarly or Wordtune, are intended for writers who need help, say because they’re writing in a second language or with a deadline looming over their heads. But according to research on the history of assistive technology, many of these tools' core features—like advanced spellcheck and grammar rules—trace their roots back to specialized software like Writer’s Workbench (from the late 1970s) or early versions of Kurzweil 3000, which were specifically created for users who have dysgraphia, dyslexia, or other cognitive processing difficulties. Ever have autocorrect correct your use of “there/their/they’re”? That’s AI, originally developed to address homophone confusion, a common difficulty for people with cognitive and language-based learning disabilities. And that smartphone feature that offers suggestions for your next word? Predictive text owes its origins to Roy Feinson, who coded the “software” behind it in the 1980s to more easily communicate with deaf people over the phone. It’s useful not only for people with motor issues (who would have trouble typing out full words), but also had an impact for people with intellectual disabilities (suggesting words to use, thereby partly bearing the cognitive load for them) and people without either (who could save time composing emails or texts without using shorthand). 

Speech Transcription

Have you ever turned on the transcription service in a video call? That’s helpful for people who have hearing difficulties (they may not be able to hear your words clearly, or at all, but they can see them on the screen), and for people like me who conduct interviews for articles (copy-pasting a transcript is a huge timesaver compared to transcribing myself). And, surprise: that’s powered by AI. Even someone with full hearing can become “situationally deaf” in a loud environment (a packed sports bar, for example).

AI Vision

There are AI-powered apps that can identify things detected by your phone’s camera. If you’re blind or low-vision and need to know what’s on a menu, what the label on a prescription says, or have some other visual data identified, there’s an AI-powered app for that. If you’re sighted and are trying to watch your calorie intake while eating out, there’s an app for that, too. The latter one wouldn’t be possible without the former. A blind user of these apps (such as Be My AI) can have an alternative to asking a person for assistance, increasing their autonomy in a world that’s not designed for the disabled. Microsoft’s Seeing AI gives blind people audio descriptions of things like text, emotions, colors and their surroundings.

When Curbs Aren’t Cut

City infrastructure can use AI to identify “micro obstacles”, like areas where a curb isn’t cut, a barrier set up for construction work, or an incline that a manual wheelchair can’t navigate, so that they know what needs fixing. The “curb cut” side of this is that fixing this infrastructure helps anyone carrying something on wheels, like a delivery driver with a heavy cart, a parent with a double stroller, or a traveler with a heavy suitcase.

Conclusion

As the above examples demonstrate, what’s good for the disabled minority is good for the abled majority. AI’s no exception; it’s the 21st-century curb cut. Tech companies’ main motivation may be to turn a profit, but they’re making the world more inclusive by the day.

AI can do some great things, but it’s made by people, and nobody’s perfect. Curb cuts are simple and static—just cut a slope into a sidewalk; it’ll remain unchanged if you walk away—whereas AI is vastly more complex and fluid. But when we get AI right, we’re not just helping people on the “margins” of society. We’re helping the center, too.

Part II in this series will explore what happens when the goal of using AI to help the disabled community meets this reality, and what steps we can take to mitigate the “growing pains.” Stay tuned! If you have any feedback, feel free to send it to mitchb@incight.org!

ABOUT the AUTHOR:

Mitch Blatt is the Editor-in-Chief of The Understanding. He graduated from Linfield College with degrees in Creative Writing and Psychology in 2018, and has been working for INCIGHT as The Understanding’s Editor-in-Chief since 2019. He lives with cerebral palsy and OCD, and doesn’t let either stand in the way of living a good life.