Picture someone at their desk, composing an email. They've been at it for forty minutes. Not because the message is complex — it's three paragraphs about a project update. But every sentence feels like a trap. Too direct and they'll be called blunt. Too casual and the intent gets lost. They rewrite the opening line for the sixth time, wondering if the tone sounds angry when they mean it to sound neutral.

Then they paste it into ChatGPT: "I am neurodivergent and I am struggling to understand this email, can you explain it to me in steps?"

Anne Brown, a late-diagnosed autistic manager, calls this transformative. She spent hours crafting descriptive emails just to avoid misinterpretation. Now a $20 monthly subscription to ChatGPT, in her words, "has changed the way I live in so many aspects."

Twenty dollars a month. No disclosure to HR. No accommodation request. No manager who has to approve anything. Just a tool that meets her where she is.

This is the part of the AI conversation almost nobody is having.

The Scale We Don't Talk About

One in five people globally could be categorised as neurodivergent — roughly 15–20% of the population. More prevalent than diabetes. More than severe mobility restrictions. More than asthma. If you lead a team of twenty, three or four of them think, process, and work differently from the way your office was designed to operate.

ADHD. Dyslexia. Autism. Dyspraxia. Not edge cases. They're the fabric of every team I've ever worked with — whether anyone said so or not.

And in Singapore, awareness is catching up fast. Search interest for "neurodiversity" increased 60x between 2019 and 2024, making Singapore the 8th-highest globally for the term over a twenty-year analysis. People are paying attention.

But attention and action aren't the same thing. A March 2025 white paper from BCG and SG Enable surveyed neurodivergent employees in Singapore and found that 62% report being misjudged as having a poor attitude. Fifty-five percent sense fear or isolation at work because colleagues become distant. Fifty-three percent cite a lack of accommodative policies.

These aren't stories about people who can't do the work. They're stories about workplaces that weren't designed for how they do it.

The Disclosure Trap

Here's the cycle that makes neurodiversity in the workplace so stubbornly difficult.

Only 49% of neurodivergent employees in Singapore disclose their conditions to employers. More than half stay silent. They have good reasons: fear of social stigma, fear of career repercussions, and 75% prefer voluntary disclosure over mandatory policies.

Sunil Chandrasekhar of BCG Singapore put it precisely: "Without disclosure, employers don't act, and without action, stigma persists."

This is the trap. Not the people. Not the employers, most of whom genuinely want to do better. The trap is the mechanism itself — a system that requires you to declare your difference before it offers to accommodate it, in a culture where declaring that difference carries real professional risk.

So here's what I think matters most about AI in this context, and it's not about productivity metrics or efficiency gains.

AI breaks the disclosure trap.

For the first time, neurodivergent professionals have access to personalised accommodations they can set up themselves. Privately. Without paperwork. Without a diagnosis code on an HR form. Without anyone's permission.

AI as Translator of Ability

Dr. Cornelia C. Walther, a senior fellow at Wharton and Harvard, framed this beautifully: "AI can serve as a sort of translator, not of language, but of ability."

That metaphor is precise. Before AI, the unwritten rule of professional life was: have the idea and perform the idea in a neurotypical way. Your solution to the problem had to arrive in the right format, at the right pace, with the right social wrapper. If you couldn't produce that wrapper — if the email took forty minutes, if the meeting notes were chaotic, if the presentation slides didn't quite flow the way people expected — then the idea itself was discounted. Not on its merits. On its packaging.

AI handles the packaging.

Kim Akers, chief operations officer for Microsoft's commercial business, has ADHD, dyslexia, and dysgraphia. She once turned down reading at her brother's wedding because she couldn't reliably read last names. Tasks like reading, writing, and organising information required extra effort long before AI existed. When she first used an AI writing assistant — feeding in what she was trying to communicate and getting back polished prose in seconds — she said it simply: "that was a game changer."

The cost of masking weighs on this. Colm McNamee, a neurodiversity activist, puts it bluntly: "I pay a price for masking my neurodivergence. It comes in waves of exhaustion. It's like putting on a mask to make other people feel more comfortable."

AI doesn't eliminate masking. But it transfers the burden from the person to the tool. The idea still comes from you. The creative leap, the pattern you spotted, the connection nobody else made — that's all human. AI just handles the translation into whatever format the workplace expects.

Maitreya Shah, Technology Policy Director at the American Association of People with Disabilities, named what sits beneath the convenience: "That feeling of agency, of being able to do things independently, with AI helping you without involving family members or caregivers — all of that feels very transformative."

Agency. Independence. Privacy. These aren't productivity features. They're dignity.

The Tools People Actually Use

What strikes me about the real-world accounts is how unglamorous the applications are. No specialised "neurodivergent AI" products with shiny marketing. Just general-purpose tools bent to deeply personal needs.

Leon Furze, an author and consultant who is AuDHD (both autistic and ADHD), records voice memos on five-hour drives, then has AI transcribe and sort them by feasibility into his task manager. He writes blog posts verbally while running, and AI formats them in ten to fifteen minutes. Before conferences, he uses AI to generate "social stories" — step-by-step guides for navigating the social expectations of events that many people find intuitive but he finds overwhelming.

As Furze himself notes: "Nobody is selling 'AI generated social stories for adults who find conferences overwhelming' or 'capture your manic car ideas before they disappear forever.' The genuine assistive value I get from these tools is deeply personal, cobbled together from my own patterns and needs."

That's the point. The most valuable neurodivergent AI tools aren't the ones labelled for neurodivergent users. They're general-purpose tools creatively applied to specific friction points that only the person experiencing them fully understands.

The data backs this up. EY's global survey of 300 employees with disabilities and neurodivergence found that 85% say generative AI creates a more inclusive workplace. Eighty-eight percent report being more productive. Sixty-eight percent say it reduced work anxieties. And 80% say it helps them play to their strengths.

Hiren Shukla, who founded EY's neurodiversity programme and himself has ADHD and dyslexia, ran a Copilot sprint with neurodivergent teams that generated 60–80 process improvement suggestions. His takeaway flips the narrative: "It's not just AI helping neurodivergence. It's the power of neurodivergence maximizing the use of Copilot."

That reframe matters. Not AI compensating for deficits. AI amplifying strengths that were always there — constrained by systems not designed for them.

The Curb-Cut Effect

There's a concept in accessibility called the curb-cut effect. When cities cut ramps into curbs for wheelchair users, everyone benefited — parents with pushchairs, delivery workers with trolleys, travellers with suitcases. Features designed for disability became features everyone uses.

AI is following the same trajectory. Spell-check started as an accessibility feature. So did predictive text. Meeting transcription. Voice dictation. Goblin Tools — which breaks overwhelming tasks into manageable steps — was built with ADHD users in mind, but ask any overwhelmed project manager if they'd find that useful.

When EY's research says 85% of neurodivergent employees find AI more inclusive, that's not just a neurodiversity statistic. It's a preview of how AI changes work for everyone. The neurodivergent users are simply the ones who notice the difference first — because for them, the gap between "before AI" and "after AI" is measured in hours of effort, not minutes.

Singapore's Moment

Singapore is better positioned for this than most places realise.

SG Enable has built genuine infrastructure — the Enabling Mark accreditation framework now covers 257 companies. The Open Door Programme provides grants for job redesign. The national target is a 40% employment rate for persons with disabilities by EMP2030. Deutsche Bank runs a Neurodiversity@Work programme here in partnership with the Autism Resource Centre. Visa's "Viable" Employee Resource Group has 380+ active members across Asia-Pacific.

But there's a gap. I haven't found a single AI-specific neurodiversity programme in Singapore. The formal programmes address accommodations through traditional channels — manager training, job redesign, workplace adjustments. All necessary. All valuable. All requiring disclosure.

Meanwhile, the 51% who don't disclose aren't reached by any of these programmes. They are, however, entirely reachable by a $20/month AI subscription.

If you're a tech leader in Singapore reading this — and I know many of you are — here's the question worth sitting with. Your neurodiversity programmes, if you have them, serve the people brave enough to disclose. What are you doing for the other half? The ones who won't tell you, because the professional risk calculus doesn't work out in their favour?

You don't need to build a programme for them. You might just need to normalise the tools they're already using. Make AI assistants standard-issue. Frame them as productivity tools, not accommodations. Let people use them without justifying why. The accommodation that requires no disclosure is the one that reaches everyone.

Edward Chew of SG Enable captured the aspiration well: "What is needed most is a shift from sympathy to empathy. With the right roles, environments and mindsets, individuals from diverse backgrounds don't just fit in — they flourish."

What This Is Really About

I keep coming back to Anne Brown and her $20 subscription. Not because the amount matters — because of what it represents. For years, she spent hours on emails that took colleagues minutes. Not because she couldn't communicate, but because the expected format didn't match how her mind works. The accommodation she needed wasn't expensive. Wasn't complicated. Wasn't disruptive. It just didn't exist yet.

Now it does.

This is AI at its most quietly radical. Not generating art or writing code or predicting markets. Just sitting between a person and a workplace that wasn't built for them, translating — not language, but ability — so the ideas come through clearly.

One in five people. In every team, in every company, in this city. Some have disclosed and found support. Many haven't. For the ones still spending hours on three-paragraph emails, masking at a price that comes in waves of exhaustion — the tools are here now.

Twenty dollars a month. No permission required.