The Future of AI: Accessiblity

This piece is a part of a three part series on the Future of AI. This three part series covers a wide range from accessibility gains, the data ecology crisis, & AI psychopathology.

In my late twenties, I was diagnosed with a tremor. It’s a tremor that I’ve seen get worse over the years, and it’s one that I expect will continue to get worse. I tried to do everything I could to manage it. Limiting caffeine. Getting more sleep. Living a healthier lifestyle but the idea that I might one day lose the ability to use my hands the way that I do was daunting. It was scary, and it came out of nowhere. I thought I just had shaky hands, and suddenly I was looking at a progressive diagnosis.

In the years since, a lot of my anxiety about the future has quietly drifted away. Not because following my doctor’s advice worked to an immense degree, but because of something else entirely. The push for productivity gains from AI builders. The drive to automate, to remove friction, to make things easier. That push has very quietly solved problems I thought I would one day have to face.

Quiet Accessibility Through Automation

I don’t worry about how I’ll code in the future anymore. I can speak into an AI agent and have it code alongside me. I use my expertise while it does the heavy lifting. I don’t worry about building slide decks either. I can ask an AI to write a script that generates them exactly how I want. It saves me from having to drag and drop text boxes with shaky hands, which is far more frustrating than most people realise.

This very loud push toward productivity, the thing AI systems are marketed on most aggressively, has quietly unlocked a huge amount of accessibility and I don’t think people are paying enough attention to that. So much of the discourse focuses on job displacement. How many roles will disappear. How much labour will be removed from the market. Those concerns are real, but they dominate the conversation to the point that we miss something else entirely. For people with physical limitations, disabilities, or chronic conditions, these tools increase the ability to live life on your own terms in ways that simply weren’t possible before.

The Curb Cut Effect

 As I was preparing the speech this piece originally came from, this section very much sat in the “good.” The pattern we’re seeing with AI accessibility reminds me of curb cuts on pavements. They were introduced to help wheelchair users navigate cities safely but the knock on effects benefited everyone. Parents with strollers. Delivery workers with trolleys. Travellers with luggage. An accommodation designed for a specific group turned out to be useful for almost everyone.

AI tools are following the same pattern. Voice interfaces built for hands free convenience help people who can’t use their hands at all. Automation designed to save time also reduces motor strain. Screen readers improved by machine learning benefit both blind and sighted users. Try listening to an article while cooking dinner. Or using ‘Read Aloud’ while studying so you can follow along without staring at a screen. If you look at something like AI powered browsing, imagine what it means for someone to navigate the web for the first time. Not just having reviews read aloud, but having images described. Charts interpreted. Full product pages understood. Completing a purchase without needing to see the screen at all. 

There’s something genuinely profound in that. Historically, accessibility has been bolted on as an afterthought. Separate modes. Special versions. Accommodations that mark you as different. But when accessibility emerges from the core design of automation, it’s no longer a special case. It’s just the product.

I’m very aware that this is a hopeful framing in a complicated moment. When enterprises hire us to design AI systems, they do so to reduce man hours. To automate processes. To cut costs. AI is used to make workflows easier, but it’s also used to eliminate roles entirely. The same automation that gives me hope is taking away livelihoods. I’ve written about those risks elsewhere and I won’t pretend they don’t exist. 

The technology that may let me keep working is also the technology that makes certain jobs obsolete. My personal relief doesn’t cancel out someone else’s unemployment. Both things can be true at the same time. The displacement is real. It demands policy attention, serious investment in transitions, and a fundamental rethink of how we distribute the gains from automation but the accessibility benefits are also real. And they’re not trivial. For millions of people with disabilities, chronic conditions, or age related limitations, AI tools are expanding what’s possible. Those gains shouldn’t be dismissed just because they’re tangled up with genuine harms. I don’t have a clean answer for how to resolve that tension. I’m not sure anyone does yet.

Beyond Productivity: Autonomy

Here’s where I want to reframe the conversation. Most discourse around AI focuses on productivity. How much faster we can work. How much more we can produce. How many tasks we can automate. Those are valid questions, but they’re not the ones that matter most to me. What matters is autonomy. 

When I think about my future with a progressive tremor, I’m not primarily worried about being less productive. I’m worried about being dependent. Needing help with things I used to do alone. The slow erosion of self sufficiency that comes with physical limitation. The AI tools I’ve described don’t just make me faster. They make me more independent. They reduce the gap between what I can do alone and what requires assistance. They extend the period of my life where I can remain self reliant. That kind of value is harder to measure. Harder to monetise. Harder to justify in a pitch deck. But for the people who need it, it matters more than efficiency gains ever will. I think about the future where buttoning a shirt becomes difficult. About the small dignities that disappear when your body stops cooperating. And about the tools that might help me hold on to those dignities just a little longer.

Social Connection and Stigma

I’m not here to evangelise. I know there are people with disabilities far more limiting than what I experience now or likely ever will but there’s another dimension that doesn’t get discussed enough. Social connection.

Disability is isolating. When you need accommodations that aren’t always available, when workflows don’t match your pace, you start to withdraw. You stop attending conferences because the logistics are exhausting. You stop collaborating because synchronous work doesn’t accommodate you. You drift to the margins. AI tools reduce that isolation in subtle but powerful ways. Remote work makes location less important. Asynchronous communication supported by transcription and summarisation lets you participate at your own pace. Voice interfaces allow contribution without announcing limitation.

 And yet, there’s still a strong vilification of AI use. A belief that work produced with AI assistance is lower quality or less legitimate. When I use AI to transcribe, it often introduces formatting quirks or sentence structures that weren’t in my original speech, but the meaning is still mine. The voice is still mine. That vilification, the idea that using these tools somehow cheapens contribution, needs to stop. Especially when those tools allow people to participate on equal footing with their peers.

AI tools are not a miracle cure for accessibility challenges. They’re not a substitute for policy, universal design, or recognising accessibility as a civil right. The broader risks of AI still exist. Displacement. Concentration of power. Misuse. None of that disappears because accessibility improves.

What I’m describing is a partial and imperfect story. A set of tools that happen to help people like me, even though they weren’t designed for us. A side effect worth noticing without celebrating uncritically. I started this piece with a diagnosis that scared me. Not because of what it meant immediately, but because of what it suggested about my future. What I’ve learned since is that the future is more accommodating than I feared. Not because anyone planned it that way, but because the logic of automation aligns with my needs. The push to remove manual friction from systems is quietly shaping the world into something more accessible.

I don’t know what my hands will be capable of in ten or fifteen years. I don’t know how far the tremor will progress. But I do know that the tools available to me are improving faster than my condition is declining. When we talk about the future of AI, we tend to focus on intelligence or efficiency. I think the more important question is how much dignity it affords. How much independence it preserves. How much human agency it expands. By those metrics, cautiously and without the hype, I’m hopeful. Not because AI is magic. Not because it will solve everything. But because it’s already solving some things quietly, without fanfare, and in ways that might let me keep doing what I love a little longer.

That’s enough reason, I think, to keep building.