AI and Neurodiversity: Why AI Is Not Neutral for Neurodivergent Minds

Most AI conversation happens at 30,000 feet. Productivity gains. Competitive advantage. The future of work.

AI and Neurodiversity in the workplace: What rarely gets discussed is what AI actually does inside the minds of the people using it, and how those effects are not evenly distributed.

Because they aren't...

...but before I start - a disclaimer.

I'm not a clinical psychologist and nor do I pretend to be. This is very much from the heart as someone who has come to understand much more about neurodivergence over the last 12 months. It is written from a deeply personal perspective and one that I experience every day. I appreciate this may be controversial; I sincerely mean no offence to anyone. It is from my own lived experience.

On with the blog:

For neurodivergent professionals, particularly those with ADHD or who are autistic, AI is neither the salvation nor the threat that mainstream coverage suggests. It is something more complex, more personal, and frankly, more consequential.

This blog is for SME founders, leaders, and teammates who are neurodivergent themselves, or who work alongside those who are. It helps explain 'Why AI and Neurodiversity Require Different Leadership Approaches'.

The way your business adopts AI will either work with how your people's minds function, or quietly work against them.

The Two Default Narratives (And Why Both Fall Short)

The public conversation has mostly landed in two places.

"AI is a godsend for ADHD and ASD."

Partially true. AI can offload working memory, structure ambiguous tasks, draft emails that navigate neurotypical social conventions, and reduce the cognitive tax of decisions that feel trivial to others but aren't.

"AI risks overwhelming neurodivergent users."

Also partially true. Infinite tools, inconsistent outputs, dopamine-driven interaction loops, and the pressure to adopt fast can deepen the very patterns that already cost neurodivergent people dearly.

Both are right. Neither is the whole picture.

The real answer depends almost entirely on how AI is introduced. And that is where Talisman's foundational principle — problem before tools — stops being a methodology and becomes something closer to a duty of care.

The ADHD Reality: Where AI Can Quietly Accelerate Burnout

A caveat first: ADHD presents differently in every person, and no short piece can do justice to that variation. What follows is a pattern — common, but not universal.

ADHD is not a focus deficit. It is a focus regulation difference — an interest-based, novelty-sensitive, dopamine-driven attentional system. I speak from my own personal experience.

Now consider what AI is:

• An infinite novelty engine

• Instant micro-rewards on every prompt

• Unlimited tools, models, techniques, optimisations

• Permission to rabbit-hole indefinitely under the banner of "productivity"

  
  

For a neurotypical professional, this is a distraction.

For an ADHD professional, this can be a trap.

The pattern repeats itself in SMEs across the country: an ADHD founder discovers AI, experiences a genuine productivity lift in week one, then spends the next six weeks optimising prompts, testing every new model release, building half-finished automations, and watching their actual deliverables drift.

Not because they are undisciplined. Because the tool is doing exactly what the ADHD brain is wired to chase — novelty, variation, reward.

And on the other side of that hyperfocus cycle is something every ADHD professional knows intimately: the crash. The burnout. The quiet shame spiral of "why can't I just finish things?"

AI did not cause that pattern. But AI can feed it faster, and with more sophistication, than almost any tool in recent memory.

This matters because the symptom — stalled work — is easy to misread as a motivation problem.

It is not. It is a scaffolding problem. And scaffolding is something leaders can build.

The Autistic Reality: Predictability, Masking, and the Cost of Inconsistency

Autism exists on a vast spectrum, and any generalisation is, by definition, incomplete. With that in mind, autistic cognition often works best with:

• Clear rules

• Predictable systems

• Explicit success criteria

• Low ambiguity

• Manageable sensory and cognitive load

  
  

Large-language-model AI violates several of these simultaneously.

The same prompt produces different outputs. A model that behaved reliably yesterday behaves differently today after a silent update. Interfaces prioritise novelty and "magic" over determinism and control. Conversational AI performs empathy but cannot be trusted to maintain consistency across sessions.

For an autistic professional, this is not merely frustrating. It can be genuinely destabilising, in ways neurotypical colleagues may never see, because the visible behaviour often looks like disengagement or resistance, when the internal experience is something much closer to unsafe.

There is a second layer that matters even more.

Many autistic professionals have spent their entire working lives masking; that is, performing neurotypical communication norms at enormous cognitive and emotional cost. AI can ease that tax by drafting emails, softening tone, translating directness into expected corporate cadence.

That is a real gain. It is also a real risk.

If AI becomes a mask amplifier, a tool that helps an autistic professional pass more convincingly as neurotypical in every interaction, the underlying fatigue does not go away. It compounds. The mask becomes tighter, not lighter. The rest days become longer. The burnout becomes harder to explain.

AI can reduce masking effort, or deepen masking dependency. Which of the two it becomes is almost never a question of the tool itself. It is a question of the culture it sits inside.

Why "Problem Before Tools" Is More Than Methodology

Talisman's core principle sounds deceptively simple: define the problem before you choose the tool.

For most businesses, this is about avoiding waste, not buying tools that don't fit, not building automations that solve nothing, not mistaking activity for progress.

For neurodivergent professionals, it is something deeper.

For ADHD minds, "problem before tools" is protective scaffolding

The ADHD brain struggles not with effort, but with prioritisation and initiation. When the entry point to AI is a defined problem with a clear success criterion, the novelty-chasing pattern is contained. You are not "exploring AI." You are solving a specific, scoped thing. The dopamine loop has a finish line.

This isn't a workaround, it is alignment. Problem-first thinking externalises the executive function that ADHD makes unreliable. The tool becomes a means, not a playground. Burnout risk drops because the work ends when the problem is solved, not when the novelty runs out.

For autistic minds, "problem before tools" is clarity-producing

Autistic cognition often thrives on explicit frameworks and defined scope. Problem-first thinking provides exactly that, clear inputs, clear outputs, clear constraints. Ambiguity is resolved before the tool ever enters the room.

It also answers the deeper question of why a tool is being used at all. Using AI to solve a business problem is categorically different from using AI to mask better. When the purpose is explicit and external (not a negotiation with your own identity) the tool stops being a compromise and becomes a lever.

In both cases, the same principle does different work. It contains where ADHD overflows. It clarifies where ASD needs clarity. Not by accident; because it respects something both experiences share - that is the need for the world to be less ambiguous, less noisy, and more defined than it naturally is.

The Truth About Neurodivergent AI Adoption

Here is something the mainstream narrative rarely says:

Many of the most sophisticated AI users in SMEs right now are neurodivergent.

ADHD founders who have used AI to finally externalise the mental load they have carried their entire careers. Autistic operators who have systematised processes no one else had the patience to decompose. Dyslexic leaders who have found their written voice for the first time. People who have spent decades developing the exact compensatory skills — constraint design, systems thinking, explicit self-awareness — that AI rewards.

The risk is not that neurodivergent professionals fall behind in the AI era.

The risk is that the surrounding culture of AI adoption (hype-driven, tool-first, always-more) burns them out before their real contribution is seen.

That is a leadership question, not a tool question.

What This Means for SME Leaders

If you lead a team that includes neurodivergent people, and statistically, you do, here is what matters:

• Don't optimise for tool adoption. Optimise for problem clarity. The more defined the problems your team is solving, the more safely AI enters your operating rhythm.

• Don't mistake hyperfocus for sustainable productivity. An ADHD teammate on fire with a new tool is not automatically a win. It is a signal to check in.

• Don't confuse AI-assisted masking with inclusion. If your autistic colleagues are using AI to "fit in" more, ask whether the environment could change so they did not have to.

• Don't lead with "try AI and see what happens." That is the worst possible framing for neurodivergent minds — all noise, no scaffolding.

  
  

Lead instead with: "Here is a specific problem. Here is what done looks like. Let's use AI where it fits."

That is not just good strategy. It is good leadership.

Final Thought

AI is not a neutral addition to any workplace. It is a psychological and operational intervention and one that lands differently depending on the mind receiving it.

For neurodivergent professionals, the difference between AI as a tool and AI as a trap is almost entirely upstream of the software itself. It lives in the way it is introduced, framed, and bounded.

"Problem before tools" is often described as a business discipline.

It is also, quietly and significantly, an act of care.