It's simple: much of the writing characteristics people associate with AI are also found in authentic autistic communication.
Autistic people with lower support needs often use formal language, precise vocabulary, an unusually consistent tone. When people use these characters as "proof" of AI, they risk treating differences in communication as suspicious.
This is just the tip of the iceberg, though. Because for someAutistic people with higher support needs, they don't even have the ability to mask. Thus, they *need* computerized tools to help them communicate. This can often sound machine-like.
There's a further problem.
Sometimes autistic people *do* use AI for accessibility. For example, for arranging their thoughts. Or perceiving tone. Or reducing the cognitive load for executive function.
"But I'm autistic and don't use AI!"—some might say.
I'm glad you don't. However, autism is not uniform. It exists as a *spectrum*. What is easy for some is profoundly difficult for someone else.
Now I'm sure there are some people who believe, "AI is totally evil—so screw any kind of accommodation."
You're entitled to have your opinion on AI. And go ahead, criticize the companies developing it. I'm not arguing that the environmental costs, labour practices, and ethics are not problematic.
But disliking AI *still* doesn't eliminate the accessibility needs of people who benefit from it.
Yet, no matter how much you dislike AI, there's still a *huge* difference between criticizing it and policing how *humans* are permitted to communicate.
Because if your defence of human authenticity requires autistic people to prove they're human, something has gone seriously wrong.
Three years ago, I was more amenable to anti-AI sentiment.
But my attitude has changed drastically.
Because what I've generally found is the anti-AI crowd tends to also be the neurotypical supremacist crowd who demands conformity to *their* particular manner of expressing themselves.