Danny Townley was diagnosed with dyslexia during childhood. He describes experiencing a “massive fog cloud” between his brain and his fingers, which made typing on a computer extremely difficult.
“There’s a barrier in-between my thoughts and the keyboard. I can’t write from an empty page, it’s impossible,” he told upstart.
Townley began using AI platforms to convert speech to text a few years ago, when AI software became more readily available. As a LinkedIn marketing and visual branding strategist, his work depends on his ability to effectively translate his thoughts into written and digital form.
“I use it for practically everything,” he says. “Emails, proposals, LinkedIn posts, anything that needs to be written, I use AI.”
Many are uncomfortable with the introduction of AI into everyday life. It has created a plethora of humanitarian issues, like deepfakes, copyright concerns and the spread of misinformation. Although, alongside these negatives, AI has also brought new assistive technologies to the world, and to the disabled communities particularly. With more accurate speech-to-text, voice recognition and vision assistance, AI is giving people with disabilities new ways to communicate, work and participate in their day-to-day life.
Much like Townley, there are millions of people across the world who are impacted by dyslexia. Kevin Murphy, founder of Lexy, an online literacy tutoring program for individuals with dyslexia, was also diagnosed with the learning disorder as a child. Growing up in Ireland, Murphy was taught Gaelic throughout school, a language that is notoriously phonetically complex.
“It’s very easy to diagnose students who are struggling with it [learning Gaelic] as having dyslexia,” he tells upstart. “So, I was given extracurricular help.”
Murphy was given a computer to use educational technology aimed at assisting his dyslexia. And from there, he went on to study computer science at university and create Lexy, which uses AI to create tailored reading and literacy programs that are personalised for each user. Using AI, the software is able to analyse specific areas where the user needs assistance and tailor the tutoring to become increasingly harder as the user progresses.
Murphy believes assistive AI will only improve as the software becomes cheaper, smarter and more readily available, creating more opportunity for AI to assist people with dyslexia and other disabilities. And as AI collects more data from users, it will be able to apply the data in different ways, and continually improve.
“AI has been a field of story for 70 years. You know it’s only over the last two that’s really kind of blown up, I think it just creates a lot of opportunities for better services.”
Legal scholar and disability rights expert Piers Gooding says assistive AI is not just making tasks easier for people with disabilities but is fundamentally reshaping how they live their life. He offers examples of two technologies that have already helped many.
“If we were talking about screen reader software or voice-to-text software, I’d say it’s a story of great success,” he tells upstart.
Another area where AI has created immediate benefits is in technology for those who are blind or have low vision, such as built-in smartphone tools and apps like Be My Eyes and Aira. These can, within seconds, read a street sign, identify a face or describe a room. These apps use AI technology alongside human assistance to support people with low vision. The AI can identify objects it’s sent without the need for human intervention.
Associate Professor Matthew Butler, head of Inclusive Technologies at Monash University, says the goal of such technology is to return agency to users, but warns it cannot go unchecked.
“They feature the ability to bring a human into the loop to validate or confirm and then it’s up to an individual’s tolerance for risk… what levels of trust they give to the AI,” tells upstart.
That trust can be fragile. A glitch at a busy crossing or an incorrect description can put users at risk.
Martin Israel, 79, has just three per cent vision in one eye due to retinitis pigmentosa, a degenerative eye disease that gradually damages the retina and will eventually result in complete blindness. He has tried advanced tools such as OrCam glasses, which can read books aloud and recognise objects he programs into them. He also experimented with handheld obstacle-detection devices that use sound cues to help him orient himself in his surroundings. Despite the availability of high-tech vision aids, Israel prefers to use what he knows works for him.
“I felt that really it wasn’t any benefit,” he tells upstart. “I can hold the cane, or I can hold on somebody’s shoulder when I’ve got my cane anyway. I have a lot of difficulty now moving around without my wife assisting me.”
Professor Lauren Ayton, deputy director of the Centre for Eye Research Australia, says many people value the simplicity, reliability and companionship of canes and loved ones over high-tech devices.
“You have a huge amount of technology coming through… but long canes are great, long canes give people lots of information about their environment, they can fold them up, they can pop them in their bag when they don’t need them,” she tells upstart.
For people with low vision, comfort, safety and agency come first and they are not willing to sacrifice them for technology they can’t fully depend on. For others, AI technology can be utterly critical in staying connected to those around them.
An example of this is Hazmir Mohamadan. Mohamadan, a 42-year-old father with a background in IT, was diagnosed with Motor Neuron Disease (MND) in February of this year. MND is a name for a group of neurodegenerative diseases that damages motor nerves, eventually leading to paralysis and breathing difficulties. Since his diagnosis, Mohamadan has been exploring eye gaze and voice banking technology so he can communicate with his family, if the time comes when he can no longer talk. Eye gaze technology works by following the user’s eye movement to select what they want to say. These technologies are often used with just computer synthesised voices, much like the voice Stephen Hawking used.
Whereas voice banking is a process where people record themselves and then convert it into a computerised voice to generate speech that sounds like their own. These technologies, which fall under Augmentative and Alternative Communication (AAC) devices, have been around since the 20th century. The first electrical AAC devices were already AI but have increasingly improved for the past 15 years along with other forms of AI.
Mohamadan says that voice banking was a simple process.
“All I had to do was record myself having conversation with my wife for at least 2 to 4 mins,” he tells upstart.
“Then, I recorded a few more speeches and conversation that’s roughly about 10 mins each. I saved those files and emailed it to my speech pathologist.”
Simon Judge is the service lead for Barnsley Hospital’s Assistive Technology team and has helped provide communication aid to patients in the Yorkshire and Humber region. He remembers it taking much longer to bank someone’s voice in the past.
“The first one I did was with someone that was about to get a laryngectomy,” he tells upstart.
“I went around loads of times with a laptop, a microphone and spent a lot of time recording,” he says. “I can’t remember how many, but it was days’ worth of work and thousands of phrases.”
As the technology improves, people now have more time later in their diagnosis to use the software, something Judge says was previously an issue.
“We tend to start working with people after the point at which they should have banked their voice,” he says.
“That’s evolving because the technology is getting better. So, they kind of potentially can make voices based on disordered speech.”
Kevin Murphy is optimistic about what AI can continue to do in the accessibility space.
“All these technologies are becoming cheaper, better, more available, quicker,” he says. “It’s a really bright future where things become more real-time and more adaptive.”
For Mohamadan, at the root of it all, his motivation is to have the technology to help him remain present.
“I want my child to still hear my voice,” he says. “My wife to be able to listen to me asking for help or just simply making a joke. I still want to be present even though I am dying.”
Article: Amy Ditcham, Josh Grobbelaar, Obiora Eze and Charlie Jones.
Photo: by author
