{"id":3012,"date":"2025-10-29T13:20:28","date_gmt":"2025-10-29T12:20:28","guid":{"rendered":"https:\/\/lab.the-bunker.it\/rubrica\/do-neurodivergent-people-use-ai-more\/"},"modified":"2025-11-26T13:17:41","modified_gmt":"2025-11-26T12:17:41","slug":"do-neurodivergent-people-use-ai-more","status":"publish","type":"rubrica","link":"https:\/\/lab.the-bunker.it\/en\/rubrica\/do-neurodivergent-people-use-ai-more\/","title":{"rendered":"Do neurodivergent people use AI more?"},"content":{"rendered":"\n

I recently found out that I\u0092m neurodivergent. Many traits that, when I was younger, got me labeled as the \u0093weird\u0094 one, the \u0093crazy\u0094 one, or sometimes even the \u0093genius,\u0094 have now merged into a specific cluster of behaviors caused by the atypical structure of my brain \u0096 and understanding this has been a useless relief.
This discovery was preceded by another one: I mostly (if not almost exclusively) hang out with neurodivergent people; among them, many study or intensively use generative artificial intelligence.<\/p>\n\n\n\n

As much as the risk of confusing correlation with causation is always around the corner, I couldn\u0092t shake off a suspicion: are these technologies particularly useful for people who have a different mental setup? For a long time, I had noticed that for me they were enabling technologies. To give an example: when I was younger, one of the reasons I didn\u0092t pursue an academic career was a specific trait of mine. Developing ideas or analyses comes extremely easily to me, but refining the details like notes and bibliographies exhausts me. Those who aren\u0092t ADHD often think it\u0092s just attitude, but it isn\u0092t; when I wrote my thesis, I remember drafting the text in a few weeks, effortlessly, while I spent much more time assembling and correcting the bibliography and footnotes. In general, once I\u0092ve created the content, the details bore me to death. It\u0092s not laziness, as I was often told, but simply the way my brain works. I can\u0092t do the things that bore me and I can\u0092t stop doing the ones I\u0092m passionate about. It may sound silly, but this trait has often hindered my work in the past. And if I say \u0093in the past\u0094 it\u0092s because with a Large Language Model, bibliographies and boring details are no longer a problem.<\/p>\n\n\n\n

\u0093But AI makes mistakes! It invents books!\u0094 some professor will say; yes, that\u0092s true. However, with informed use those errors are easily fixable (just Google things or have the AI Google them for you) and the time savings are still significant. And even if it gets things wrong\u0085 well, it still does so less than I do.<\/p>\n\n\n\n

This is just one example of how this new technology has given me what I\u0092d call a pharmacological relief in many everyday tasks, work-related and not. Wondering if I was the only one came naturally. I talked about it with Alberto Puliafito and Mage De Baggis (who fall into the category of neurodivergent friends working with AI) and they confirmed my impression. My partner, who is autistic, told me the same thing, even though the benefits she finds are different from mine. In short, anecdotes among friends and colleagues piled up, and even though I\u0092ve met neurodivergent people who hate AI (though often without ever really using it), it was worth investigating. Puliafito had written about it here, prompting me to take the investigation one step further. What does the research say? In education, there\u0092s plenty, but what I wanted to understand wasn\u0092t whether they are useful for teaching neurodivergent learners (they are), but how and whether this community uses them in private and professional settings.<\/p>\n\n\n\n

\"\"\/
Image courtesy of Francesco D’Isa.<\/figcaption><\/figure>\n\n\n\n

So I did some research \u0096 or rather, several deep searches \u0096 and found some interesting results. It seems that many neurodivergent individuals report that LLMs provide a more comfortable and nonjudgmental communication environment compared to interpersonal interactions with neurotypicals. In particular, autistic people like my partner often find it easier to talk to a chatbot than a human, because AI-mediated conversation removes many of the ambiguities of social signals and offers more direct responses. A recent experimental study by Carnegie Mellon University compared the reactions of autistic adults when asking for advice from two different interlocutors: a GPT-4-based chatbot and a human counselor whom participants believed to be a second chatbot. Participants overwhelmingly preferred the real chatbot over the human disguised as an AI (Jang, 2024). Not because the AI gave better advice, but because of the interaction style: the chatbot replied quickly, explicitly, and in a structured way \u0097 for example with bullet points and clear instructions \u0097 whereas the human counselor tended to ask follow-up questions and request clarifications.<\/p>\n\n\n\n

This preference for a direct, nonjudgmental tone also appears in online discussions within neurodivergent communities. A large-scale qualitative analysis (61 neurodivergent Reddit communities) shows that many autistic users appreciate ChatGPT because it can \u0093offer advice or comfort without judgment or envy\u0094 (Carik et al., 2024). Having a conversational partner that is always available and free of social bias also helps these individuals communicate with less anxiety. For example, one autistic college student said she uses ChatGPT to practice difficult conversations (simulating dialogue scenarios with roommates), and that this gives her a sense of autonomy: she can try out what to say without involving others right away or relying on her parents (Carik et al., 2024). Others report writing a sensitive message or email and having the AI \u0093review\u0094 it first, checking whether the tone is appropriate. In short, generative AI can offer neurodivergent people a communication channel free of social pressure, where they can get feedback or advice in a neutral and predictable way. This is particularly useful in facing pragmatic challenges (tone of voice, social conventions, misunderstandings) that often make interactions with neurotypical interlocutors stressful (Richter, 2025).<\/p>\n\n\n\n

Beyond communication, generative AIs are also used by neurodivergent people as tools for organization, brainstorming, and productivity \u0097 especially by those with attention-related profiles such as ADHD, which is exactly my case.<\/p>\n\n\n\n

\"\"\/
Image courtesy of Francesco D’Isa.<\/figcaption><\/figure>\n\n\n\n

Individuals with ADHD often struggle with time management, task structuring, and maintaining attention; here, AI can act as a \u0093cognitive amplifier.\u0094 A 2025 study on online communities shows that users with ADHD widely use AI chatbots as productivity tools: about 26% of the analyzed discussions in ADHD forums concerned the use of LLMs to organize activities, improve focus, or optimize workflow (Carik et al., 2024). In academic contexts, university students with ADHD, dyslexia, or other learning differences are adopting ChatGPT and similar tools to help them with writing tasks: a survey of 124 students with disabilities (mostly with writing difficulties due to ADHD, dyslexia, dyspraxia, autism) found that the most widely used generative tools were chatbots (especially ChatGPT) and rewriting applications (Zhao et al., 2025). They use them in every phase of writing \u0097 from brainstorming, to generating paragraphs, to grammar revisions \u0097 confirming that these tools can bridge some executive and attentional gaps during academic work.<\/p>\n\n\n\n

Of course, the use of AI is not without issues, and neurodivergent users themselves are aware of them. In the aforementioned survey, for example, many students with ADHD or learning disabilities acknowledged the practical benefits of ChatGPT but expressed concerns about possible inaccuracies and risks to academic integrity (accidental plagiarism, excessive reliance on the tool) (Zhao et al., 2025). Qualitative data also highlights a certain frustration in interacting with AI: users with ADHD report difficulties in refining prompts and \u0093making the LLM truly understand what I mean.\u0094 In fact, one study notes that over half the time, ADHD users complain about frustrating prompting and perceive \u0093neurotypical\u0094 biases in AI responses \u0097 for instance, models insisting on verbose explanations or styles poorly suited to the direct processing style they prefer (Carik et al., 2024). This suggests that, from a usability perspective, there is room to improve AI adaptation to these cognitive needs. In general, however, generative assistants are largely seen as allies in managing daily activities.<\/p>\n\n\n\n

We\u0092re still in the early days and more research will come, but it seems my impression wasn\u0092t entirely baseless. On a personal level, it also explains my frustration when someone complains about the \u0093good old days\u0094 before AI and how it will supposedly ruin creativity. As far as I\u0092m concerned, I have no doubt that it helps me and boosts my creativity, even just by freeing me from boring tasks. Forgive me, those who want to go back to pen and paper, but the \u0093good old days\u0094 when a minute of bibliography work weighed on me like an hour\u0085 I absolutely do not miss them.<\/p>\n\n\n\n

Bibliography<\/strong><\/h5>\n\n\n\n

Carik, Buse, Kaike Ping, Xiaohan Ding, and Eugenia H. Rho. 2024. \u0093Exploring Large Language Models Through a Neurodivergent Lens: Use, Challenges, Community-Driven Workarounds, and Concerns.\u0094 arXiv<\/i> (preprint), October 8, 2024. https:\/\/arxiv.org\/abs\/2410.06336<\/a>.<\/p>\n\n\n\n

Jang, JiWoong (Joon), Sanika Moharana, Patrick Carrington, and Andrew Begel. 2024. \u0093\u0091It\u0092s the Only Thing I Can Trust\u0092: Envisioning Large Language Model Use by Autistic Workers for Communication Assistance.\u0094 In Proceedings of the CHI Conference on Human Factors in Computing Systems (CHI \u009224)<\/i>, May 11\u009616, 2024, Honolulu, HI. New York: ACM. https:\/\/doi.org\/10.1145\/3613904.3642894.<\/p>\n\n\n\n

Richter, Hani. 2025. \u0093\u0091It\u0092s the Most Empathetic Voice in My Life\u0092: How AI Is Transforming the Lives of Neurodivergent People.\u0094 Reuters<\/i>, July 26, 2025. https:\/\/www.reuters.com\/lifestyle\/its-most-empathetic-voice-my-life-how-ai-is-transforming-lives-neurodivergent-2025-07-26\/<\/a>. Reuters<\/a><\/p>\n\n\n\n

Zhao, X., Cox, A., & Xuanning, C. (2025). The Use of Generative AI by Students with Disabilities in Higher Education<\/i>. The Internet and Higher Education, 66, 101014. https:\/\/doi.org\/10.1016\/j.iheduc.2025.101014<\/a><\/p>\n\n\n\n

Francesco D’Isa<\/h5>\n\n\n\n

Francesco D’Isa, trained as a philosopher and digital artist, has exhibited his works internationally in galleries and contemporary art centers. He debuted with the graphic novel I. (Nottetempo, 2011) and has since published essays and novels with renowned publishers such as Hoepli, effequ, Tunu\u00e9, and Newton Compton. His notable works include the novel La Stanza di Therese (Tunu\u00e9, 2017) and the philosophical essay L’assurda evidenza (Edizioni Tlon, 2022). Most recently, he released the graphic novel \u0093Sunyata with Eris Edizioni in 2023. Francesco serves as the editorial director for the cultural magazine L’Indiscreto and contributes writings and illustrations to various magazines, both in Italy and abroad. He teaches Philosophy at the Lorenzo de Medici Institute (Florence) and Illustration and Contemporary Plastic Techniques at LABA (Brescia).<\/p>\n","protected":false},"author":8,"featured_media":2171,"template":"","meta":{"_acf_changed":false,"inline_featured_image":false},"rubriche_category":[51],"class_list":["post-3012","rubrica","type-rubrica","status-publish","has-post-thumbnail","hentry","rubriche_category-la-rivoluzione-algoritmica"],"acf":[],"_links":{"self":[{"href":"https:\/\/lab.the-bunker.it\/en\/wp-json\/wp\/v2\/rubrica\/3012","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/lab.the-bunker.it\/en\/wp-json\/wp\/v2\/rubrica"}],"about":[{"href":"https:\/\/lab.the-bunker.it\/en\/wp-json\/wp\/v2\/types\/rubrica"}],"author":[{"embeddable":true,"href":"https:\/\/lab.the-bunker.it\/en\/wp-json\/wp\/v2\/users\/8"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/lab.the-bunker.it\/en\/wp-json\/wp\/v2\/media\/2171"}],"wp:attachment":[{"href":"https:\/\/lab.the-bunker.it\/en\/wp-json\/wp\/v2\/media?parent=3012"}],"wp:term":[{"taxonomy":"rubriche_category","embeddable":true,"href":"https:\/\/lab.the-bunker.it\/en\/wp-json\/wp\/v2\/rubriche_category?post=3012"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}