This blog was authored with the assistance of real and imagined friends and colleagues.
I’m not alone in the growing interest in how neurodivergent people are using Generative AI (GenAI) on their own terms. Rather than these tools being imposed in clinical or educational contexts (as interventions to “fix” or accommodate), many people are using them in self-directed, pragmatic and creative ways that align with lived experience. This shift reflects something bigger than mere tool adoption – it’s a reframing of what support can look like.
A seismic shift for support
GenAI is increasingly becoming a form of personal infrastructure, quietly embedded in day-to-day life. Among 2025 use case trends, people aren’t just using GenAI for productivity, but for therapy, companionship, life organisation and even exploring personal purpose. That’s a seismic shift – not just in what AI does, but in who defines value, how trust is earned and what counts as support.
Two areas stand out to me as especially resonant for neurodivergent users:
- Communication without judgment – Many neurodivergent (and neurotypical) people turn to GenAI to rehearse conversations, decode confusing social cues or simply talk through a thought without fear of being misunderstood or judged. When social interactions have historically involved effortful decoding or fear of misinterpretation, the ability to externalise your thinking without resistance is a quiet revolution.
- Executive function as cognitive unburdening – AI also provides real, low-friction support with executive function challenges. This includes help with breaking down tasks, writing clearer emails, scaffolding complex projects and building/sustaining routines. These are often framed as productivity features in workplace or educational settings, but for many neurodivergent people they offer something deeper: cognitive relief.
A systemic challenge and a radical opportunity
Crucially, these aren’t just stories of GenAI assisting users; they’re stories of people appropriating tools to build autonomy. This flips the traditional narrative. Neurodivergent users are not passive recipients of assistive technology – they’re active agents in shaping how GenAI gets used for executive functioning, emotional processing and communication.
If individuals are finding more usable, effective support through free GenAI tools than through the labyrinth of diagnosis-driven, rationed institutional systems, then the question arises: are our models of educational and clinical support still fit for purpose?
This question gains urgency when we consider equity. As one colleague noted, students from more advantaged backgrounds often have easier access to personalised, human support. If GenAI can offer some of this universally, it may reduce the stigma attached to seeking traditional help. For learners marginalised by systems that don’t understand or accommodate their needs, the AI assistant is not just a tool – it’s a refuge.
…but not without risks
Of course, the promise comes with serious caveats. Four key risks are:
- Over-reliance – GenAI excels at giving structured answers and gentle encouragement –sometimes too much. Studies have raised concerns about AI sycophancy, where the model always affirms the user, even when gentle challenge might be healthier or more productive. If we rely too heavily on AI for social rehearsal or emotional reassurance, we might inadvertently erode confidence or independence in other settings.
- Commercialisation – these tools are largely created, owned, and operated by private companies, whose incentives may not align with public good or long-term care. If we cede too much of the emotional or cognitive support ecosystem to for-profit entities, we risk creating dependencies that aren’t sustainable.
- Masking and homogeneity – relying on GenAI to help navigate neurotypical expectations could reinforce ‘masking’ (the exhausting practice of hiding or suppressing neurodivergent traits to fit in). If AI tools help us appear more socially acceptable or “normal,” are we creating new pressures to conform and lessening the acceptance of difference?
- The digital divide – even as we celebrate GenAi tools’ empowering potential, they are not equally accessible and we must stay alert to who is left out (e.g. those without stable internet, up-to-date devices, or digital literacy). If GenAI becomes an informal layer of support infrastructure, there’s a risk it will reproduce the same inequities of formal systems.
So, what now?
For GenAI to help neurodivergent people function more freely, we should focus on who gets to shape them. Neurodivergent users should be included not only as test subjects or case studies, but as designers, decision-makers and ethicists in the development of this technology.
And we must ask what our institutions are doing to evolve. If support is now ambient, flexible and user-led, our educational and clinical systems must learn to support this shift. We can’t hold onto outdated models of gate-kept services and one-size-fits-all interventions when people are clearly finding better fits on their own.
A final question lingers for me: when people with lower support needs use AI to navigate gaps in human support, does this supplement, replace or highlight the failures of traditional systems? And might these self-guided tools, in their own quiet way, be a call to reimagine support from the ground up?
Coming up in Part 2 of this 3-part series: GenAI and the art of procedural wayfinding for neurodiverse learners
This is so great Shaun! I’ve supported a few Accessibility students with Co-Pilot to help manage their studies. Looking forward to Part 2
Thank you so much for the lovely feedback, Amanda! You can now find pt. 2 here https://educationexpress.uts.edu.au/blog/2025/05/20/neurodivergence-and-genai-part-2-procedural-wayfinding/