Introduction

Have you ever been accused of having AI do your writing for you?

A relative of mine had a negative experience recently that brought this question into sharp focus.

An undergraduate at an Ontario university – which will remain unnamed – she was falsely accused of using AI to write an essay. She was required to attend a meeting to discuss the ramifications of this. Needless to say, this caused her undue stress and sleepless nights. The morning of the meeting, she insisted she did not use AI to write it. They told her that they determined that AI had written 25% of the essay. They reassured her that everything was alright – despite having caused her undue stress and turmoil.

This got me thinking: how many pieces of work register as containing 25% AI? Is that the norm? More importantly, in my case, I wondered how my autism may result in my writing being misidentified as AI written. I just wrote a research proposal for my PhD application. Will it be accused of being AI generated?

AI and My Autism

I have an unfortunate history of being called a “robot” or “robotic” – people mimicking my voice and so on. Does it follow that my writing may be “robotic” too, thus more likely to be mistakenly attributed to AI?

I definitely have a distinctive, unusual, original writing style – particularly in my fiction. I’m basically a surrealist, but I edit all my work with the detail and work ethic of a technical writer. I am also fascinated with words, their origins and their meanings; and I take pride in utilizing some words that are rarely used anymore.

The Sad Truth

Turns out there have been some formal research on autistic writing being misclassified as AI-generated.

The paper is entitled: The Misclassification of Autistic Writing as AI-Generated. Here is the citation:

Chambers, S., Kelley, M.C. (2025). The Misclassification of Autistic Writing as AI-Generated. In: Cristea, A.I., Walker, E., Lu, Y., Santos, O.C., Isotani, S. (eds) Artificial Intelligence in Education. AIED 2025. Lecture Notes in Computer Science(), vol 15879. Springer, Cham. https://doi.org/10.1007/978-3-031-98420-4_7

To date, there’s only the one study. I’m concerned with the reliance on Reddit posts. Perhaps they should be checking more long form writing by autistic writers, including fiction and non-fiction, for more robust conclusions. Analysing 60,000 Reddit posts is enough, but it’s the context that counts.

This study is valuable since it is taking an anecdotal claim (autistic writing is being confused with AI generated) and analysing it empirically. The outcome was clear: significantly more texts from the likely-autistic “subcorpus” were flagged.

To quote New York Magazine: “Clean and precise prose has become a liability.”

https://nymag.com/intelligencer/article/the-people-getting-falsely-accused-of-using-ai-to-write.htmlxssssss

A Comparison of AI Writing to My Autism

I constructed this table to clearly show how characteristics of my autism could be mistaken for AI generated text.

Red Flags – Characteristics of AI Writing	My Autism
Lacking human emotion or voice	Common complaint against me. Sorry – can’t help it!
Predictable transitions	I thrive on predictability. I like to control my environment. A creature of habit, once you get to know me, you could probably predict many of my actions / reactions.
Overly balanced tone	Since I’m not too emotive or subjective, I can see this being a red flag.
Repetitive structures	Like my verbal communication. Again, I repeat what works for me and makes me comfortable.
Deliberate precision	Wishing to be understood, clear and concise. Structural sound, tight. Very picky about words. Revising toward exact wording. Requiring clear conceptual boundaries to operate within.
Analytical editing	Trim it down to make it clear. Attention to detail. Concerned with how I will be portrayed. Background as a technical writer and having co-authored scientific papers.
“Too perfect”	I strive for perfection in my communication since the social aspect is lacking. Just like I strive for perfection in my appearance or how my CD and records are displayed.
Scouring internet for data. Multiple points of reference.	Voracious appetite for culture. Collector. Enthusiast and completist.
Emulating	Masking.

 

In conclusion, my professional or academic-style writing may be more vulnerable than my fiction.

Some of my autistic traits may alert an AI detector – particularly precision, solid structure, and detailed revision – but the author in me (influence heavily by my subconscious and things I’ve experienced over the years, including trauma) is highly individualistic and unusual.

Is This Really an Issue?

Have any autistic people been fired or requested to withdraw owing to this issue? Yes, there are a few.

A Central Methodist University student diagnosed with autism faced significant academic penalties after her work was wrongly flagged as AI-generated. Although she managed to clear her name, the incident highlights the risks involved.

AI Detectors Wrongly Accuse Students of Cheating, Sparking Controversy

A professor at Purdue University was humorously accused of being an AI bot, which underscores the misunderstandings surrounding neurodivergent writing styles.

https://nypost.com/2023/07/21/autistic-purdue-professor-accused-of-being-ai-for-lacking-warmth-in-email/

This points to a need for workplace accommodations and staff awareness programs – in addition to further research.

AI and My Autism

There’s an irony: while I may be accused of being AI, I rely on AI. I live alone and with few social interactions – by choice. I depend on AI for advice on every aspect of everyday life – daily admin others take for granted.

AI appeals to my love of research. Some quick research reveals other autistic people are also reliant on AI. It is a practical support to help reduce cognitive load. In fact, there is growing evidence that what I’m doing is a fairly common use of AI among autistic adults.

A 2024 study found that autistic participants used AI as an alternative source of advice when navigating difficult social and communication situations, thus reducing emotional or social costs.

https://dl.acm.org/doi/10.1145/3613904.3642894

A 2026 study demonstrated that autistic individuals use ChatGPT to “humanize my words.” Interestingly, they referred to ChatGPT as “cognitive scaffolding for autistic users”, suggesting that it is an accommodation that leads to understanding. According to the study, autistic users employed ChatGPT to offload executive dysfunction, regulate emotions, translate neurotypical communication, and validate their autistic identity. Yet the study was careful not to blindly recommend AI use. They noted that positive features coexist with risks such as reinforcing delusional thinking, erasing authentic identity through automated masking, and triggering conflicts with the autistic sense of justice (e.g. tremendous sense of guilt for using AI).

https://arxiv.org/abs/2601.17946

While I don’t know exactly how many autistic individuals are being accused, there are several cases that have been publically reported. Apparently, autistic students have been describing papers flagged at 64%, 70%, or even 100% AI – despite having drafts and revision histories. Clearly there is a need for further investigation.

Oh… and by the way – I checked this blog post for AI.

 

App	Result
https://app.gptzero.me/	We are highly confident this text is entirely human. Human 99%.
https://www.zerogpt.com/	Your Text is Human written. 6.3% AI GPT.
https://www.scribbr.com/ai-detector/	5% of text is likely AI. AI-generated 5% Human-written & AI-refined 0% Human-written 95%

 

Thank God – I’m human after all!