Crip Doulaing, Neurodivergence and AI: A conversation with Heather McCain
Learning and engaging with Heather McCain, Founder and Executive Director of Live Educate Transform Society (LET'S) on disability justice, lived expertise, access, accountability, and community control.
In this piece, Heather McCain generously shares knowledge with Lara McLachlan on Crip Doulaing, what AI can make possible for neurodivergent people and the risks and harm already occurring, and why disabled and neurodivergent communities need meaningful power in shaping what comes next.
Please tell us about yourself and LET'S
I'm Heather McCain, Founder and Executive Director of Live Educate Transform Society (LET'S). At LET'S, we centre the expertise of people who are 2SLGBTQIA+, disabled, neurodivergent, and otherwise equity-denied, and we utilize that lived knowledge to transform education, policy, and community practices.
Everything LET'S does is rooted in community. We are guided by the knowledge, needs, priorities, and dreams of those most impacted. By staying accountable to community, listening, adapting, and co creating, we ensure LET'S is not simply delivering services but helping to build and strengthen networks of care and resistance that can extend beyond any single workshop or project.
You describe yourself as a Crip Doula. What does being a Crip Doula mean to you, and how has your own experience shaped that work?
I am proudest to be called a Crip Doula. This is a disability justice term created by organizer Stacey Park Milbern to describe the ways disabled people support and mentor newly disabled people in learning disabled skills. and navigating a world that wasn't built with us in mind. This title means the most to me as it was gifted by community members who have felt the positive effects of my work.
For me, being a Crip Doula means using the pain, frustration, and stress of my own experiences and transforming it into support to ensure other people don't have to go through what I did. I focus on creating spaces and practices where people can grieve, re-imagine, and reclaim their lives, while honouring and celebrating the incredible organizers, advocates, and activists who have fought, and continue to fight, for a better and more possible future.
Read more about Disability Doulaing.
Can you tell us more about disability justice and how came into this work?
Disability justice calls for cross movement work, collective access, and collective liberation, which means we need multiple entry points: education, systems change, relationship building, and community support.
When I was growing up, I didn't have the word for my asexual, neurodivergent, queer, or trans identities and I didn't recognize my disabilities as disabilities. There was a lack of awareness and knowledge, and it deeply impacted my life and, eventually, my health. I lived in a world where I constantly navigated systems, spaces, and people that didn't take me or my identities into account. There was so much about myself I knew in my bones, but I didn't have language to express it or know how to start the conversation.
That relentless stress of masking, translating myself for others, and enduring environments and people who refused to see me took a serious toll on my physical health and mental wellbeing. Over time, the effort of suppressing my needs and identities, and trying to make myself understandable to myself and to people who were not willing or interested in understanding, pushed my body into survival mode.
My nervous system was doing crisis management every day, anticipating harm, rehearsing scripts, bracing for misunderstanding, and my body eventually responded in the only way it could, through chronic illness and autoimmune conditions. The weight of being unseen, misread, and pathologized didn't stay abstract. It showed up as pain, exhaustion, and a kind of deep burnout that impacted every area of my life and continues to impact me in my 40s.
That embodied impact is part of why I talk about disability justice as not just an idea but a necessity. It's why I fight so hard for spaces where people don't have to mask to be safe, where our bodies are believed, and where our identities are understood as valid and whole.
Read more about disability justice principles.
What led you to create a justice-centred AI ethics guide?
Artificial intelligence is a conversation happening in many non-profit and community spaces. However, the conversations rarely centre disability justice, Indigenous sovereignty, or intersectionality in meaningful ways.
In response, LET'S created our Justice-Centred AI Ethics Guide and companion guide specific to research. My goal was to support organizations and individuals in understanding both the potential benefits and the very real harms of AI.
I observed harmful patterns in these organizational discussions. AI was framed as either harmful or transformative, with little nuance or room for complexity.
People were being shamed for using AI as an access tool. Others were judged for not adopting it quickly enough. These reactions limit meaningful dialogue. We need deeper conversations grounded in care, complexity, and accountability.
AI is not neutral. It is built by people and companies and reflects the priorities, values, biases, and power structures of those who design and utilize it. Without intentional intervention, AI systems will continue to reinforce and increase systemic inequities, including ableism, racism, transism, colonialism, and other "isms".
Justice-centred AI asks us to do more than minimize harm. It challenges us to actively build systems that support and uphold dignity, access, and self-determination, especially for communities who have historically and are currently bearing the brunt of technological experimentation and surveillance.
AI has the potential to support accessibility and reduce barriers. However, this will only happen with intentional action, accountability, care, and consent. We cannot be passive in how AI develops. The stakes are too high.
Can you tell us more about the concept of Data Sovereignty?
Many people interact with AI every day, but do not know the term "data sovereignty" or have a clear understanding of what it means and how it shapes these systems. Part of this gap is intentional. Conversations about data sovereignty are often led by academics, policymakers, and technology companies in ways that make the concept feel abstract, technical, or out of reach. Legal framing and industry jargon can create barriers, making it difficult to see how these issues directly affect people's everyday lives. In reality, data sovereignty is about power, consent, and whose knowledge and experiences are used to build AI. As a result, something that is fundamentally about people's rights and autonomy gets framed as something only experts can define or influence.
Business interests also play a role in keeping the concept inaccessible. Many AI systems depend on large-scale data extraction, and a broader public understanding of data sovereignty could lead to increased demands for consent, resistance to harmful practices, or refusal to participate altogether. When people are not given the tools or language to understand what is happening with their data, it becomes easier for decisions to be made without meaningful consent, accountability, or transparency.
Making data sovereignty more accessible is not just about education; it is about shifting power. When people can clearly understand how their data is used, they are better positioned to question, refuse, and advocate for approaches that respect their rights, knowledge, and lived experiences.
For governments, data sovereignty often refers to data being governed by the laws of the region where it is stored and processed. While this can shape protections related to privacy, surveillance, and cross-border data sharing, it does not automatically ensure that equity-denied communities are respected or protected.
Communities, especially Indigenous, disabled, and other equity-denied groups, have the right to decide if, how, and by whom their data is collected, used, and shared. This includes the right to refuse extraction entirely. That applies to medical records, education data, social service files, and online activity.
Can you tell us more about the concept of Algorithmic Harm?
Algorithmic harm refers to the ways AI and automated decision-making systems leave people or communities worse off than they would have been without them. For example, a hiring algorithm may rank neurodivergent candidates lower because their communication styles or work histories do not match narrow definitions of "ideal" workers. Without the system, those candidates might still have had a chance; with it, they are filtered out automatically, reinforcing exclusion and economic precarity.
Similarly, predictive policing tools often direct increased surveillance toward communities that have already been over-policed, including Black and Indigenous communities, disabled people, and people living in poverty. Because these systems rely on historical data, they reproduce and intensify existing harms rather than improving safety.
Taken together, data sovereignty and algorithmic harm are about power: who decides what is normal, whose data is extracted and used, and who bears the consequences when systems fail. Justice-centred approaches insist that those most affected must have real control over whether and how these systems are used, including the ability to refuse participation when the risks outweigh the benefits.
Please tell us more about the ways AI is being used now.
AI is already shaping decisions that affect disabled people's access to income, housing, healthcare, and employment, often without clear notice or consent. This includes automated benefits reviews that flag people for reassessment or cutoffs and hiring tools that filter out candidates based on gaps in employment or non-linear career paths.
In healthcare and social services, algorithmic triage and risk-scoring systems can influence who receives care, how quickly, and at what level of support. Disabled people may be assessed as "lower priority" based on assumptions about quality of life or cost of care or flagged as "high risk" in ways that increase surveillance rather than support. These systems are often difficult to challenge, especially when decisions appear automated or are not clearly explained.
At the same time, AI is deeply embedded in many of the tools disabled people rely on for access, communication, and autonomy. AI-powered screen readers, captioning and transcription tools, predictive text, and speech recognition support blind and low-vision communities, Deaf/deaf and hard-of-hearing people, and those with speech and cognitive disabilities to navigate information, participate in education and work, and communicate on their own terms. Smart home systems, mobility devices, and other assistive technologies increasingly use AI to adapt to a person's needs and environment.
From a justice-informed perspective, a core risk of AI is that it is trained on historical data shaped by systemic inequities. This means AI systems often learn and reproduce patterns of racism, ableism, transim, anti-immigration, and criminalization, and can scale those harms more quickly and widely than human decision-making alone.
A second risk is the lack of transparency and accountability. People are often not told when AI is being used, what data it relies on, or how decisions are made. This makes it difficult to question outcomes, challenge errors, or give meaningful consent, undermining due process and self-determination.
These risks show up differently across communities but follow similar patterns of exclusion and control. For neurodivergent people, AI systems are frequently built around narrow, normative assumptions about communication, behavior, productivity, and "risk," which means we're more likely to be flagged as non compliant, unsafe, or less employable.
AI is not neutral infrastructure; it reflects the priorities of the systems that build and deploy it. For disabled people, the issue is not only whether AI is used, but how, where, and under whose control.
The same technology can be used to deny benefits or to generate accurate captions; the difference lies in whose interests shape the design, governance, and everyday use of these systems.
Do think AI can be beneficial or even liberatory for neurodivergent people?
For neurodivergent people, AI is only liberatory when it is shaped by our input and used to remove barriers, rather than to enforce or measure us against rigid, "normal" standards. When AI tools are designed without us, they tend to reproduce the same narrow expectations about communication, productivity, emotion, and behaviour that already make many spaces inaccessible.
By contrast, when neurodivergent people are involved in defining the problems, setting the goals, and testing the tools, AI can be used to create flexible structures that actually meet our needs. That might mean supporting alternative communication styles, adapting to non-linear work patterns, and/or helping manage sensory and cognitive load in ways that respect our boundaries.
As an Autistic with ADHD (attention deficit hyperactivity disorder), AI can function as a practical access tool in my day-to-day work. I use it to break down complex tasks into smaller steps, re-organize long documents into clearer structures, and turn tangled notes or partial thoughts into outlines I can act on. Because I often move into hyperfocus and gather a lot of information very quickly, AI is useful for helping me sort, prioritize, and sequence that information so it becomes usable rather than overwhelming.
As a small non-profit led by disabled and neurodivergent people, AI is used by some of us as an access and sustainability tool. We use it for drafting, formatting, organizing, planning, customizing, and creating resources in ways that respect our health, energy, and fluctuating capacity while supporting our process.
In addition to being Autistic and having ADHD, I am a disabled, queer, and trans person. My energy levels and health are deeply impacted by the systemic oppression I encountered and continue to encounter.
AI helps redistribute some of the cognitive and administrative load so I am not constantly running on depletion. It can take on parts of the work that are repetitive or bureaucratic, like structuring documents, reformatting content, or generating variations of materials, so that more of my limited energy can go toward analysis, relationship-building, and community care.
At the same time, I am intentional about how I use AI so that I am not outsourcing my thinking or amplifying existing harms. I treat AI as a drafting and organizing partner, not an author: I always return to the output with my own analysis, rewrite in my own language, and check that what is on the page still reflects my politics, experience, and voice. I cross-check facts against trusted sources, compare information across multiple documents, and pay careful attention to whose knowledge is being cited or erased, rather than assuming an AI-generated answer is accurate or neutral.
I also try to "train" the systems I use, as much as they allow, to align with justice-centred values rather than simply extracting from them. That can mean feeding in disability justice and neurodiversity-affirming frameworks, correcting ableist or oppressive language when it appears, and prompting in ways that center collective care, accessibility, and consent. When a response misses context or reinforces harmful assumptions, I name that, adjust how I am using the tool, and sometimes discard the output entirely.
When 2SLGBTQIA+, disabled, neurodivergent, and other equity-denied folks have real control over how AI is designed and governed, these tools can redistribute labour and create more time and capacity for community care.
In my own work and within our organization, AI is most helpful when it supports our access needs, structuring, customizing, and de-cluttering information, while we maintain oversight over content, verify facts, and stay accountable to our communities.
You describe the guide as a living document. How can the neurodiversity community contribute to what it becomes?
Calling the guide a living document is an invitation. It will only be as strong and relevant as the communities who shape it. Naming it as a living document signals that it is meant to evolve alongside technology, politics, and our own understanding, rather than stay frozen in the moment when it was first created.
Neurodivergent people can contribute by sharing concrete examples of how AI is affecting them (positively and negatively) in workplaces, schools, health care, benefits systems, and online spaces. Those stories help ground the guide in real experiences instead of abstract scenarios and make it more useful for others navigating similar systems.
Folks can also help identify gaps: terms that need clearer definitions, contexts that are missing, or intersections (like climate justice, Indigenous sovereignty, mad politics, trans justice, or incarceration) where more collective thinking is needed. Naming where the guide feels thin, incomplete, or too general is itself a contribution.
That participation can look like sending feedback, co-creating case studies, participating in workshops or focus groups, or developing community-owned guidelines that we then integrate back into the document. It might also mean contributing questions we have not yet answered, or proposing new sections based on emerging technologies and harms.
My goal is for the guide to remain responsive to real-world conditions, so that it can keep supporting disabled and neurodivergent people to navigate AI with both caution and agency. Over time, I hope it reflects not just my perspective, but a wider tapestry of lived experience, analysis, and strategies generated by the communities most affected.
Thank you so much to Heather for this insight and leadership. I cannot overstate the importance of this discussion. For NCF, AI is a vital area in which to build partnerships and create new opportunities to advance equity for the community.
Stay tuned as we continue to explore topics of AI, neurodivergence, support, autonomy, power and learning with partners.
References
Cole, D. (2026, July 17). Black prisoners are assigned harsher living conditions in Ontario jails—thanks to AI. The Breach. https://breachmedia.ca/black-prisoners-are-assigned-harsher-living-conditions-in-ontario-jails-thanks-to-ai/
Dawson, S. (2024). Stacey Park Milbern. National Women’s History Museum. https://www.womenshistory.org/education-resources/biographies/stacey-park-milbern
Disability Justice Network of BC. (n.d.). Disability Justice Alliance: A collective protocol for global solidarity and systemic liberation.https://djnbc.ca/disability-justice-alliance/
Live Educate Transform Society. (n.d.). AI ethics guides. https://www.connectwithlets.org/ai-ethics-guides/
Mithani, J. (2023, July 26). Newly disabled people aren’t given a ‘how-to’ guide. Disability doulas are closing those gaps. The 19th. https://19thnews.org/2023/07/disability-doulas-support-newly-disabled-people/
