AI for Communication Accessibility

TL;DR:

I debated posting this here. This is a safe space for me, but based on the responses I got in other groups, I think it is an important and time disability access discussion…

I’m not a robot. I’m not a cheater. I’m a disabled, neurodivergent professional who uses AI to communicate more clearly, because for years, I couldn’t. This isn’t about shortcuts. It’s about access.

Yes, I know the environmental cost. I care. But I’m tired of being judged by people who think they can hear AI in my writing and assume that means it isn’t mine. It is mine. And I’m done apologizing for using the tools that finally let me be heard.

—————————

Someone here recently told me that something I wrote “reeks of AI.” They didn’t say exactly that, but memory fails me. And it wasn’t with curiosity. It felt like a quiet way of saying my writing wasn’t real. That it didn’t count.

And it’s not the first time something like that has happened. Some time ago, I authored a template physician’s letter designed to support families in pursuing insurance coverage for needed accessibility devices. I drew on clinical research and spent considerable time refining the language to ensure it was both medically accurate and persuasive within insurance frameworks. I used AI as a collaborative tool to help organize and streamline the writing process, but the clinical reasoning, structure, and content were entirely my own.

Later, a colleague shared a version of that letter she had adapted for her own use with an insurance company. She expressed pride in having written it independently, without the influence of AI. Her version reflected effort and a clear sense of investment in the outcome.

However, it was unmistakably based on the original letter I had written. Despite this, the way it was framed implied that her contribution was inherently more authentic, while mine was relegated as less valid as I had used collaborative tools to improve the outcome.

But it still stung. It felt like everything I had done, the hours of thinking, shaping, building, had been erased. Like needing support made the final result less authentic. Like what I create, or what I help others create, doesn’t really count.

I understand this “imposter syndrome” feeling more than most. I didn’t write anything of my own until the end of high school. My father used to redline every paper I brought home. Every sentence. Every paragraph. He was trying to help, but all I learned was how to shut down. I stopped writing because I was afraid of being torn apart. Not because I didn’t have something to say.

So when someone implies that what I write isn’t mine, or that using help makes it less real, I feel it. Deeply.

Sometimes, yes, AI adds things I don’t love. Sometimes it sounds off. But I edit. I shape it. I make it mine. And while I do that, I think about the cost. The electricity. The water. The footprint. I don’t feel great about it. But I also think about all the years I was locked out of my own words.

One example? Those long dashes. Em dashes. AI loves them. It litters them throughout everything. I’ve trained myself to spot them instantly, because I know the second someone sees one, they’ll assume the whole piece was written by a machine. I spend a ridiculous amount of time just trying to clean them out. Not because I care that much about punctuation, but because I know how quickly people use those tiny signals to write me off.

And honestly, it reminds me of how people listen to autistic speech. That slightly robotic cadence. The pauses that don’t follow the usual rhythm. The flatness. Not everyone has it, but some do, and when you do, people notice. I see it in my son. I hear it in some of the kids I work with. It’s a giveaway. A tell. And once people hear it, they decide. You’re different. You’re probably autistic. You’re marked as disabled. And that mark follows you whether you want it to or not.

That’s what an em dash feels like to me now. Just one more thing that outs me as someone not speaking in the right rhythm. Not writing the way people expect.

And it’s the same thing when I tell people I’m AuDHD. I do mask, but I also believe in being transparent. Especially with my patients. I think it matters. Some time back, I was in a meeting with professionals, a few parents, and some patients. I was open about being neurodivergent because I think honesty builds trust. I thought the conversation had gone really well. Later, a speech pathologist reached out to me privately. And the second I saw her message, I just knew what was coming.

She mentioned she had some thoughts about how I communicated during that conversation.

She didn’t say anything harsh, and she may have had helpful intentions. We didn’t end up talking further after the meeting, so I don’t know exactly what she meant, but something in her message made me brace for the kind of feedback I’ve received many times before. It’s possible she was referring to my pacing, or how I brought up certain points, or something about the flow of the conversation. Maybe it had nothing to do with my neurodivergent identity. But when you communicate the way I do, you learn to read between the lines.

She struck me as very neurotypical in how she communicated. She tended to stick closely to one topic at a time, and some of the ideas I brought up seemed unfamiliar to her, even though others in the room seemed to follow them easily. I could tell that something about my approach threw her off, even if she didn’t say so directly.

That’s what this is. The dashes. The tone. The pacing. The diagnosis. All of it. Little tells that I’m not quite doing it right.

The stupid dashes keep outing me. I take them out again and again, but they sneak back in. And the second someone sees one, they stop listening to what I’m saying and start judging how I said it.

Honestly, it’s not just the dashes. It’s all of it. The critiques. The feedback. The well-meaning corrections. The professional suggestions. The rebukes that are framed as help. We spend years absorbing those messages, being shaped by them, flinching before someone even says something because we already know what’s coming.

It’s not just the disability. It’s the constant reaction to the disability. That’s what makes us hypervigilant. That’s what makes us anxious. That’s what makes us feel like we’re never quite enough. Not because we can’t communicate, but because we’ve been told so many times, in so many small ways, that the way we communicate is wrong.

It’s just like how I was told in graduate school that I would never be a good clinician.

Yet here I am.

And it’s not like people wouldn’t have figured it out anyway. I interrupt. I info-dump. I struggle with small talk. I don’t always pace myself the way people expect. So if I don’t say anything, I’m just seen as rude. Abrasive. Overbearing. And if I do say something, I’m seen as unprofessional. There is no version that makes people comfortable except silence. And I’m done with silence.

Of course, it doesn’t come off that way to other neurodivergent people. They just get it. We have similar conversation styles for similar reasons. It feels normal to them. Comfortable, even. That’s the double empathy problem right there. It’s not that I’m failing to connect. It’s that the people I’m talking to are reading me through a lens I never chose.

And yes, I’ve seen AI hallucinate. I even wrote a song about it, called “Citation Hallucination.”

https://suno.com/song/64311d1d-c4c3-48d2-a092-be4b239d425b

At one point, I asked AI to fix the punctuation in a blog post, just to get rid of some em dashes. I had already fact-checked every citation myself. But the tool decided there was a gap in the logic and quietly filled it in. It added twelve completely fake citations, spread across more than sixty different blog posts. I had to go through every single one of them with a fine-tooth comb, searching line by line to figure out what was real and what had been slipped in.

That wasn’t convenience. That wasn’t passive. That was labor.

And it was a side effect. Just like the medications so many of us in the disability community have to take. That metaphor was mine. I’m the one who made the connection. AI helped expand it, cleaned up the wording, yes. But the core idea came from lived experience, not the tool.

We take the pills that help us focus, but they raise our blood pressure. We take the meds that stabilize our mood, but now we can’t sleep. We rely on tools to function, and the tools come with a cost.

So does that sound like laziness to you? Does it sound like I’m just letting a machine do the work?

Because what I did wasn’t cutting corners. It was repairing damage I didn’t cause. It was three days of cleanup triggered by one request to fix punctuation.

Even with all that, AI forgets. It doesn’t hold the whole idea. I have to keep reminding it what I said, what I meant, where we were going. My own memory fails too. So I rebuild the scaffolding again and again.

It’s a bit like playing with a live choose-your-own-adventure that comes up with random ideas. Many fit, and others don’t. If you don’t have a driver, you’re going to steer right into the weeds.

With AI’s help, I’ve written a blog that has reached hundreds of thousands of people. Parents. Professionals. Other disabled folks. People I never could have reached alone. That blog is helping people feel seen. It’s helping others understand. None of that would have happened without AI. Not because it wrote for me, but because it finally let me write.

Meanwhile, I’m raising two neurodivergent kids. And yes, we use AI in our homeschool too. But not without boundaries. My son once copied a Wikipedia paragraph instead of writing his own draft. So I don’t let them use AI without supervision. But I do let them create with it.

My daughter wrote a song using ChatGPT and Suno about how growing up is hard, and how metamorphosis isn’t like it is in the picture books. It’s not The Very Hungry Caterpillar. It’s painful. It’s confusing. It’s real.

Here’s her song: Caterpillar Lies

https://suno.com/song/0a32049d-f556-411b-9b76-996baed058cb

She revised the lyrics herself. She checked the rhythm. She adjusted the rhyme. She started learning to play the melody on piano using an AI-based app. I’ve paid tutors before. It didn’t work. Now she’s learning because she cares.

She also earns points in our token economy. Real points for real effort. Reading. Cleaning. Making music. She once saved up 3,500 points, about $150 worth, and used it to book a room at Great Wolf Lodge. That was her choice. Her initiative. Her accomplishment.

I couldn’t parent this way without support. These systems don’t replace me. They extend me. They let me show up when I’d otherwise fall short. And yes, they cost something. Power. Time. Money. But they also give back in ways that nothing else has.

So no, I don’t want to live in a world where people walk around with glowing numbers over their heads showing how much water they’ve used or how much their life has cost the planet. I don’t want my children or anyone else’s to be reduced to how efficient they are.

Because I’m not just using AI. I’m shaping it. I’m adapting with it. And if you think that makes my voice less real, then you’re not listening.

The wizard behind the curtain isn’t a computer. It’s me.