What Neurodivergent Thinking Can Teach Us About A.I.

Staring at a blank page. Having an idea but struggling to figure out where to start. Knowing exactly what you want to say, but spending far too much energy figuring out how to say it. 

These are all too familiar frustrations for many of us, but Madelaine Robillard, BSW’20, MSW’22, believes artificial intelligence can offer another way through them, not by doing our thinking for us, but by removing some of the friction that gets in the way.

It’s a perspective Robillard has developed from an unusual intersection of professional expertise and lived experience. A Registered Social Worker and certified sexologist, she is the founder and CEO of Revolution Therapy & Wellness, a multidisciplinary therapy and wellness collective. She is also a Canadian subject-matter expert in ethical AI integration for mental health and a neurodivergent professional herself.

"Instead of treating it as a productivity hack, I started treating it as an accommodation. When used intentionally, it falls in the same category as a noise-cancelling headset or a standing desk."

This fall, Robillard will bring those perspectives together for Alumni All-Access in Wired Differently: Working Smarter with AI as an Accessibility Tool for Neurodivergent Minds. At the heart of the conversation is an idea that applies well beyond the session and, she argues, well beyond those who identify as neurodivergent.

“I have ADHD, am autistic and gifted; defined as twice-exceptional, or 2E, which in practise means I think in systems; rapidly, sometimes messily, and with capacity that looks different moment to moment,” says Robillard.

For her, using AI effectively isn’t necessarily about doing more, faster. It’s about removing the friction between having an idea and being able to bring it to life. 

“Working with your brain means building support around how you actually think, not around what we are taught productivity should look like,” Robillard says.

From productivity hack to accommodation

Robillard runs a business, sees clients, writes and thinks in what she describes as “big systems.” Generating ideas has never been the problem; translating all of that big-picture thinking into manageable steps can be.

“The gap between what I could envision and what I could execute was what was inaccessible,” she says. As she began finding ways to use AI to bridge that gap, she stopped thinking about the technology simply as a productivity tool. “Instead of treating it as a productivity hack, I started treating it as an accommodation. When used intentionally, it falls in the same category as a noise-cancelling headset or a standing desk.”

That shift first emerged in structured and repetitive work surrounding clinical documentation. AI support didn't change the quality of Robillard’s clinical thinking; it changed how much capacity the surrounding task consumed. 

“Accommodations aren't shortcuts and they aren't about cheating,” says Robillard. “They're what lets someone's actual capability reach the outside world without the cost of masking or burnout along the way.”

The same principle can apply to more ordinary tasks. Robillard uses AI to get past a blank page, workshop and prioritize ideas, organize information, and break larger projects into manageable steps. As a direct communicator, she can also use it to rephrase something when she wants to make sure her tone is sensitive or appropriate. Rather than asking AI to complete a task, her approach is to identify the part creating the barrier and ask whether the tool can help with that piece.

“None of these are about AI generating the thinking,” she says. “They're about AI absorbing the structural labour around the thinking so there's capacity left for the parts that actually need a human.”

Robillard describes the ideal role of AI as “a really smart, intuitive companion.” It can help structure, organize, question and provide insight, but the person using it remains the expert, responsible for reviewing the result and making the decisions.

When accessibility makes things better for everyone

Although Robillard approaches these tools partly through her own neurodivergence, the barriers she describes aren't exclusive to neurodivergent people. Everyone has limited capacity. 

"Accommodations aren't shortcuts and they aren't about cheating," says Robillard. "They're what lets someone's actual capability reach the outside world without the cost of masking or burnout along the way."

Robillard points to a principle of accessible design: build for the edges and the middle benefits, too. For example, adjusting street curbs for wheelchair users also makes life easier for parents pushing strollers and workers using dollies. She sees a similar opportunity in designing ways of working around the reality that our brains don't all process information in the same way. 

“Tools built to reduce cognitive load for neurodivergent brains reduce cognitive load for everyone, because everyone has limited capacity, even if the limits are for different reasons,” she says. “The lesson isn't, ‘We should all think like neurodivergent people.’ It's that designing for real variation in how brains work produces better systems than designing for an assumed neurotypical default and treating everyone else as an exception to accommodate later.”

That perspective can challenge assumptions about productivity, too. If the right structural support removes the friction and work suddenly flows, Robillard says, the barrier may not have been a lack of discipline or capability in the first place. But there’s a danger in turning every barrier removed by technology into an expectation to produce more. 

“If reclaimed time or energy just gets filled with more tasks, that's not accommodation, it's a faster treadmill,” she says. “Use AI to develop process and capacity, not output. What you do with the energy you get back is the real measure of whether it worked.”

The human still matters

For someone whose work sits at the intersection of mental health and ethical AI, Robillard is equally interested in where the technology's usefulness ends.

That starts with privacy. She recommends understanding what happens to information before entering it into an AI tool and is particularly firm that identifying client, patient or third-party information should never be entered into a general consumer AI tool.

There are also situations where AI shouldn't be part of the interaction at all. Robillard draws the line around significant relational stakes or potentially irreversible consequences, including grief, conflict with someone you love, major ethical or clinical decisions and choices affecting another person's safety or wellbeing. 

“If getting this wrong would hurt someone, or if the value of the interaction is the relationship itself rather than the information, that's a sign to step back from the tool and toward a person,” she says.

Critical thinking matters in lower-stakes uses, too. AI can produce incorrect information and be overly agreeable. Having a tool help structure something you understand is very different from allowing it to produce material on a subject you don't know well enough to verify. The goal isn't to surrender judgment to AI, but to create more room for human judgment.

“Pick one task that quietly drains you, not your biggest problem, and see what happens when you hand the structural part of it to a tool for a week,” Robillard says. “You're the expert on your own experience, and the tool's job is to reduce friction, not to replace your judgment,”

For anyone intrigued, but intimidated by the seemingly endless array of tools, terminology and advice about crafting the “perfect” prompt, Robillard recommends starting much smaller.

“Pick one task that quietly drains you, not your biggest problem, and see what happens when you hand the structural part of it to a tool for a week,” she says.

Tell the tool what you're trying to accomplish, where you're getting stuck and what kind of support you need. You can even describe the challenge and ask how AI might help.

“You're the expert on your own experience, and the tool's job is to reduce friction, not to replace your judgment,” Robillard says.

It’s a principle Robillard will explore further during her Alumni All-Access session. Rather than beginning with everything AI can do, her approach invites a more personal starting point: How does your brain work? Where do you get stuck? And what could you do with the time, energy or capacity you might get back?