Opportunities for AI in Accessibility

I was completely moved by Joe Dolson’s current article on the crossroads of AI and availability because I found it to be both skeptical about how widespread use of AI is. In fact, I’m very skeptical of AI myself, despite my role at Microsoft as an accessibility technology strategist who helps manage the AI for Accessibility award program. AI can be used in quite productive, inclusive, and accessible ways, as well as harmful, exclusive, and harmful ways, just like with any tool. Additionally, there are a lot of uses in the subpar center as well.

I’d like you to consider this a “yes … and” piece to complement Joe’s post. I’m not trying to reject any of what he’s saying, but rather to give some context to initiatives and options where AI may produce real, positive impacts on people with disabilities. I want to take some time to talk about what’s possible in hope that we’ll get there one day. I’m no saying that there aren’t real challenges or pressing problems with AI that need to be addressed; there are.

Other text

Joe’s article spends a lot of time addressing computer-vision types ‘ ability to create alternative words. He raises a lot of legitimate points regarding the state of the world right now. And while computer-vision concepts continue to improve in the quality and complexity of information in their information, their benefits aren’t wonderful. As he rightly points out, the state of image research is currently very poor, especially for some graphic types, in large part due to the lack of context for which AI systems look at images ( which is a result of having separate “foundation” models for words analysis and picture analysis ). Today’s models aren’t trained to distinguish between images that are contextually relevant ( should probably have descriptions ) and those that are purely decorative ( couldn’t possibly need a description ) either. However, I still think there’s possible in this area.

As Joe points out, human-in-the-loop publishing of ctrl text should definitely be a factor. And if AI can intervene to provide a starting place for alt text, even if the swift might say What is this BS? That’s not correct at all … Let me try to offer a starting point— I think that’s a win.

If we can specifically teach a design to consider image usage in context, it might be able to help us more swiftly distinguish between images that are likely to be beautiful and those that are more descriptive. That will help clarify which situations require image descriptions, and it will increase authors ‘ effectiveness in making their sites more visible.

While complex images—like graphs and charts—are challenging to describe in any sort of succinct way ( even for humans ), the image example shared in the GPT4 announcement points to an interesting opportunity as well. Let’s say you came across a map that was simply the description of the chart’s name and the type of representation it was: Pie map comparing smartphone usage to have phone usage in US households earning under$ 30, 000 annually. ( That would be a pretty bad alt text for a chart because it frequently leaves many unanswered questions about the data, but let’s just assume that was the description in place. ) If your website knew that that picture was a pie graph ( because an ship model concluded this ), imagine a world where people could ask questions like these about the creative:

Would more people use smartphones or other types of phones?
How many more?
Is there a group of people that don’t fall into either of these pots?
What number is that?

For a moment, the chance to learn more about images and data in this way may be innovative for people with low vision and blindness as well as for those with different forms of color blindness, mental disabilities, and other issues. It could also be helpful in education settings to help people who can see these figures, as is, to understand the data in the figures.

What if you could request your website to make a complicated map simpler? What if you asked it to separate a single line from a range graph? What if you could request your website to transform the colors of the various lines to work better for variety of colour blindness you have? What if you asked it to switch colors in favor of patterns? That seems like a possibility given the chat-based interfaces and our current ability to manipulate images in today’s AI tools.

Now imagine a purpose-built model that could extract the information from that chart and convert it to another format. For instance, it might be able to convert that pie chart (or, better yet, a number of pie charts ) into more usable ( and useful ) formats, like spreadsheets. That would be incredible!

Matching algorithms

When Safiya Umoja Noble chose to put her book Algorithms of Oppression, she hit the nail on the head. Although her book focused on how search engines can foster racism, I believe it’s equally true that all computer models have the potential to foster conflict, prejudice, and intolerance. Whether it’s Twitter always showing you the latest tweet from a bored billionaire, YouTube sending us into a Q-hole, or Instagram warping our ideas of what natural bodies look like, we know that poorly authored and maintained algorithms are incredibly harmful. A large portion of this is attributable to the lack of diversity in those who create and shape them. However, when these platforms are built with inclusive features in mind, there is real potential for algorithm development to help people with disabilities.

Take Mentra, for example. They serve as a network of employment for people who are neurodivers. They match job seekers with potential employers using an algorithm based on more than 75 data points. On the job-seeker side of things, it considers each candidate’s strengths, their necessary and preferred workplace accommodations, environmental sensitivities, and so on. On the employer side, it takes into account each work environment, communication strategies for each job, and other factors. Mentra made the decision to change the script when it came to traditional employment websites because it was run by neurodivergent people. They use their algorithm to propose available candidates to companies, who can then connect with job seekers that they are interested in, reducing the emotional and physical labor on the job-seeker side of things.

When more people with disabilities are involved in developing algorithms, this can lower the likelihood that these algorithms will harm their communities. That’s why diverse teams are so crucial.

Imagine that a social media company’s recommendation engine was tuned to analyze who you’re following and if it was tuned to prioritize follow recommendations for people who talked about similar things but who were different in some key ways from your existing sphere of influence. For instance, if you follow a group of white men who are not white or aren’t white and who also discuss AI, it might be wise to follow those who are also disabled or who are not white. If you followed its advice, you might gain a more in-depth and nuanced understanding of what’s happening in the AI field. These same systems should also use their understanding of biases about particular communities—including, for instance, the disability community—to make sure that they aren’t recommending any of their users follow accounts that perpetuate biases against (or, worse, spewing hate toward ) those groups.

Other ways that AI can assist people with disabilities

I’m sure I could go on and on about using AI to assist people with disabilities, but I’m going to make this last section into a bit of a lightning round if I weren’t trying to put this together in between other tasks. In no particular order:

voice recognition is. Researchers like those in the Speech Accessibility Project are paying people with disabilities for their help in collecting recordings of people with atypical speech. As I type, they are currently hiring people with Parkinson’s and related conditions, and they intend to expand this list as the project develops. More people with disabilities will be able to use voice assistants, dictation software, and voice-response services, as well as to use only their voices to control computers and other devices, according to this research.
Text transformation. The most recent generation of LLMs is quite capable of changing existing text without giving off hallucinations. This is incredibly empowering for those who have cognitive disabilities and who may benefit from text summaries or simplified versions, or even text that has been prepared for bionic reading.

The importance of diverse teams and data

We must acknowledge that our differences matter. The intersections of the identities we live in have an impact on our lived experiences. These lived experiences—with all their complexities ( and joys and pain ) —are valuable inputs to the software, services, and societies that we shape. Our differences must be reflected in the data we use to develop new models, and those who provide it need to be compensated for doing so. Stronger models can be created using inclusive data sets, which lead to more equitable outcomes.

Want a model that doesn’t demean or patronize or objectify people with disabilities? Make sure that the training data includes information about disabilities written by people with a range of disabilities.

Want a non-binary language model? You may be able to use existing data sets to build a filter that can intercept and remediate ableist language before it reaches readers. Despite this, AI models won’t soon replace human copy editors when it comes to sensitivity reading.

Want a copilot for coding that provides recomprehensible recommendations after the jump? Train it on code that you know to be accessible.

I have no doubts about how dangerous AI can and will be for people today, tomorrow, and for the rest of the world. However, I also think we should acknowledge this and make thoughtful, thoughtful, and intentional changes to our approaches to AI that will reduce harm over time as well. Today, tomorrow, and well into the future.

Thanks to Kartik Sawhney for assisting me with writing this article, Ashley Bischoff for her invaluable editorial assistance, and of course Joe Dolson for the prompt.