How AI Tools Are Changing Life for Neurodivergent Workers
You are sitting in a meeting, and for the first time in your career, you are not panicking about the notes. The AI is transcribing everything. It will pull out the action items, summarise the discussion, and send you a clean document you can actually process at your own pace. The colleague next to you has no idea this changes anything for you. That is the point.
For millions of neurodivergent workers, a quiet revolution is underway. The tools that were built for productivity are turning out to be some of the most effective cognitive accommodations ever created — and they are available to everyone, no disclosure required.
How many neurodivergent workers are already using AI
The numbers are striking and still accelerating. A 2026 survey conducted by The Harris Poll for Understood.org, sampling 2,073 US adults including 614 who identified as neurodivergent, found that 78 percent of neurodivergent employees are now using AI tools at work. That is nearly twenty percentage points higher than the 59 percent adoption rate among neurotypical employees.
The gap tells a story. Neurodivergent workers are not just keeping pace with AI adoption. They are leading it — because for them, these tools are not conveniences. They are compensatory infrastructure.
A separate study commissioned by Microsoft from EY in late 2024 surveyed 316 neurodivergent and disabled employees using Microsoft Copilot across 16 organisations. The findings were consistent: 91 percent viewed Copilot as a valuable assistive technology, 88 percent reported increased productivity, and 85 percent said the tool supported a more inclusive workplace.
A UK Department for Business and Trade study in late 2025, surveying over 1,000 workers across multiple industries, found that neurodiverse workers were 25 percent more satisfied with AI assistants than their neurotypical counterparts. The people these tools help most are also the people who recognise their value most clearly.
Why AI works as a stealth accommodation
Traditional workplace accommodations require a process that most neurodivergent adults dread. You disclose. You fill out forms. You wait for approval. You become the person with the special arrangement, visible in a way that many would rather avoid.
The disclosure barrier is not imaginary. The 2026 Understood.org survey found that 70 percent of neurodivergent employees worry that disclosing their condition could negatively impact their career — up from 59 percent just two years earlier. Seventy percent also reported having no idea what accommodations they were entitled to. For neurodivergent women, the picture is worse: 22 percent who requested accommodations were denied outright, compared to 7 percent of neurotypical women.
AI tools sidestep this entire architecture. When an AI writing assistant, a meeting transcription service, or a task-management app is available to every employee as standard workplace software, a neurodivergent worker can access cognitive support without singling themselves out. The accommodation is invisible because it is universal.
Mentra, an employment platform focused on neurodivergent talent, describes this dynamic directly: AI is "secretly also an assistive technology for neurodivergent workers." The secrecy is not deception. It is dignity. The 57 percent of neurodivergent employees who say they would be more likely to disclose if employers provided specialised AI tools as standard accommodation are telling you what the real barrier has been all along — not the condition, but the cost of revealing it.
What these tools actually do for different cognitive profiles
The impact of AI tools is not uniform across all neurodivergent experiences, because the underlying cognitive differences are not uniform either. Understanding which dimensions of cognition a tool addresses explains why the same AI assistant can transform one person's workday and barely register for another.
Expression and output. For workers whose bottleneck sits between thought and written language — a pattern common in dyslexia and dysgraphia — AI writing assistants have been transformative. The EY-Microsoft study found that 80 percent of neurodivergent Copilot users received meaningful help with written communication, 71 percent achieved first drafts faster, and 61 percent reported improved ability to structure their thoughts on the page.
A 2026 study by Zhao, Chen, and Cox, published in SAGE Journals, surveyed 124 disabled students and identified three affordances of generative AI for writing: explainability (the AI could clarify why a sentence was unclear), expressibility (it helped bridge the gap between knowing what you mean and getting it onto the page), and plannability (it provided structural scaffolding before the writing began). For someone whose expression and output pathway has always been the narrowest part of their cognitive architecture, these are not luxuries. They are load-bearing supports.
Memory and sequencing. AI meeting summarisers, task managers, and scheduling tools address what is often the most taxing dimension for workers with ADHD: holding multiple threads, deadlines, and action items in working memory while simultaneously performing the actual work. The EY-Microsoft study reported that 59 percent of users found Copilot specifically helpful for memory and recall, and 55 percent for processing fine details.
Tools like Goblin Tools — a free, single-purpose AI utility designed specifically for neurodivergent users — take this further. Its "Magic ToDo" feature breaks an overwhelming task into granular steps, with a "spiciness" dial that controls how fine-grained the breakdown becomes. Tiimo, a visual daily planner built from the ground up for ADHD and autistic users, won Apple's iPhone App of the Year in 2025. These are not productivity apps repurposed for accessibility. They are accessibility tools that happen to make everyone more productive.
Attention and rhythm. Real-time transcription and captioning address a different bottleneck entirely. For workers with auditory processing differences, the gap between hearing speech and decoding its meaning in real time can consume so much attentional bandwidth that following a meeting becomes an exercise in reconstruction rather than participation. Live captions in Microsoft Teams, Google Meet, and Zoom — originally developed for deaf and hard-of-hearing users — reduce this cognitive load by providing a visual channel that runs in parallel with the auditory one. The worker with an auditory processing difference no longer has to choose between listening and understanding.
As Tara DeZao, a marketing executive with ADHD, put it: "Standing up and walking around during a meeting means that I'm not taking notes, but now AI can come in and synthesise the entire meeting into a transcript and pick out the top-level themes."
The masking tax and what AI reduces
This is where the dimensional lens becomes essential, because what AI tools are actually reducing is not just task difficulty. They are reducing the cognitive cost of masking.
Every neurodivergent worker who rereads an email four times before sending it, who builds elaborate workaround systems to simulate executive function, who avoids speaking in meetings because their processing speed does not match the conversational pace — every one of them is spending working memory on performance rather than work. The masking tax is invisible in every performance review, but it is measurable in every cognitive profile.
When an AI assistant drafts the email, the worker's cognitive resources are freed from the proofreading loop. When a meeting summary arrives automatically, the worker does not have to choose between appearing engaged and actually capturing the information. When a task manager breaks an amorphous project into steps, the executive-function gap stops being a gap at all.
The EY-Microsoft study captured this shift precisely. Participants reported that Copilot's benefits clustered around exactly the dimensions that masking most heavily taxes: written communication, memory and recall, concentration and focus, and processing fine details. The tool was not making the workers smarter. It was letting them stop spending intelligence on concealment.
"It's not just AI helping neurodivergence. It's the power of neurodivergence maximising the use of Copilot. When you harness that divergence and partner with AI, you'll see greater innovation, higher use cases, more ideation and application of AI." — Hiren Shukla, Global Neurodiversity Leader, EY
The latest wave of accessibility features
The past two years have seen the major technology platforms move beyond bolt-on accommodations toward integrated cognitive accessibility.
Apple unveiled its Accessibility Reader ahead of WWDC 2026 — a tool that creates a customised reading experience adaptable across the spectrum from dyslexia to low vision. It generates AI-powered summaries, handles complex multi-column layouts like scientific articles, and processes everything on-device, meaning no reading data leaves the user's phone. The company also introduced auto-generated subtitles for uncaptioned video content, processed locally.
Google's Gemini integration in Workspace now offers natural-sounding read-aloud with adjustable playback speeds in Google Docs, task-to-calendar scheduling, and notification minimisation during focus periods. Google partnered directly with Understood.org to incorporate neurodivergent user insights into these features.
Microsoft's Copilot continues to expand across its 365 suite, and the company marked the tenth anniversary of its Neurodiversity Program in April 2025 by releasing a best practices guide built from a decade of internal data on neurodivergent hiring and workplace support.
The pattern across all three platforms is the same: features designed for cognitive accessibility are becoming default features, not opt-in accommodations. This is the infrastructure change that matters most, because universal availability is what eliminates the disclosure barrier.
The critics are right about some things
The optimism needs tempering. Several legitimate concerns deserve attention.
The most substantive is cognitive offloading. A study of 666 participants found a significant negative relationship between frequent AI usage and critical thinking. The concern is that when AI handles the hard parts — organising, drafting, summarising — the neural pathways for those skills may atrophy over time.
For neurodivergent workers, this concern has a different texture. Cognitive offloading via AI can lower friction and scaffold working memory in ways that keep a person engaged with work they would otherwise abandon. The distinction between support and replacement depends on whether the tool preserves the worker's agency and judgment while reducing mechanical load. A meeting summary that lets you think about the content rather than scramble to capture it is not deskilling. An AI that writes your analysis for you might be.
The digital divide is real. Access to these tools is uneven. The 2026 Mentra survey found that 35 percent of neurodivergent respondents were uncomfortable with AI or had not used it at all. Cost, digital literacy, and employer willingness to provide tools all create barriers that fall hardest on the workers who need them most.
Privacy is the third concern. Neurodivergent individuals may be particularly vulnerable to the behavioural data that AI tools collect. Patterns of use — how often someone needs a task broken down, how many times they revise an email, when they struggle with focus — could, in the wrong organisational context, become evidence of the very differences the worker was trying to keep private. Many AI platforms retain user data for model training unless users actively opt out.
None of these concerns invalidates the transformation. All of them demand that it be shaped deliberately rather than left to market forces.
What this means for how you work
If you are a neurodivergent worker, the practical question is not whether to use AI tools. You probably already are. The question is whether you are using them with an understanding of which cognitive dimensions they are actually supporting.
An AI writing assistant helps most when your bottleneck is expression — the gap between what you know and what you can get onto the page. A meeting transcription tool helps most when the bottleneck is auditory processing or working memory under attentional load. A task manager with AI decomposition helps most when the bottleneck is executive function — the ability to see a large goal and generate the first step.
Understanding where your own friction actually sits — whether it is in phonemic processing, working memory, attention regulation, expression, or some combination — turns generic tool adoption into targeted cognitive support. A dimensional profile, like the one CognitionType maps across seven cognitive dimensions, can clarify which tools will make the most difference for your specific architecture and which are solving a problem you do not actually have.
The larger shift is cultural, not technological. Nathan Friedman, co-president of Understood.org, framed it clearly after the 2026 survey results: "AI is a powerful force, giving neurodivergent employees more confidence to apply for roles and disclose their differences. But this positive trend is undermined by persistent stigma, despite strong awareness of neurodiversity."
The tools are here. The stigma is not gone. And the gap between what technology makes possible and what workplaces actually support remains wide enough that 70 percent of neurodivergent employees still fear the consequences of being known.
The most important thing AI tools may ultimately do for neurodivergent workers is not summarise their meetings or proofread their emails. It is prove, at organisational scale, that different cognitive architectures produce excellent work when the environment stops penalising the difference. The accommodation was never the point. The recognition is.
CognitionType is an informational cognitive assessment, not a clinical diagnosis. If you suspect ADHD, dyslexia, or another neurodevelopmental condition, we encourage you to seek a formal evaluation from a qualified professional.