0. Why I’m interested
In college I did a couple extracurriculars that got me involved in the local disability community (Special Olympics, helping in a classroom at the high school, etc). That led me to take a yearlong class about autism. I learned a lot in that seminar and it really changed my mind and perspective.
My main conviction is that AI and VR will eventually help support autistic people in education and provide much better solutions than the current options.
1. Autism is real but not fully understood
Autism Spectrum Disorder (ASD) is listed in the DSM-5 as a neurodevelopmental disorder. ASD collapses various disorders previously referred to separately (e.g., Asperger’s, PDD-NOS, CDD).
But ASD is behaviorally diagnosed — as opposed to developmental delays caused by a distinct genetic marker, such as those related to Down syndrome. The criteria for a behavioral diagnosis are (a) deficits/differences in social communication/interaction, (b) restricted, repetitive patterns of behavior/interest/activities.
These criteria are quite broad. Of course there is some more detailed criteria and to a diagnosis. But the criteria are so broad that as to demonstrate the lack of a consensus understanding. And because it is behaviorally diagnosed — even by experts — that introduces some variation.
ASD is also partially genetically heritable — estimates of the variance explained by genetics from twin studies range from 38% to 90%.
“Real” meaning anyone can’t just say they have ASD. That doesn’t make it truly the case. My belief is that there is a true underlying condition. But because we’re diagnosing on behavior something with a complex genetic link, there’s clearly much more to learn.
2. Autism diagnoses are on the rise
ASD prevalence is increasing. Meaning that a larger and larger share of children are eventually receiving autism diagnoses. For example, the CDC reports a prevalence of 1 in 150 in 2000 rising to 1 in 36 in 2020.
Because it’s behaviorally diagnosed we don’t necessarily know why prevalence is increasing. It is some combination of (a) changes in the diagnostic criteria, (b) getting “better” at diagnosing, or (c) actual increase in the underlying base rate.
Let’s sidestep the rabbit trail of trying to disentangle those reasons.
3. Autistic people see the world differently
Many autistic people experience some level of sensory sensitivity, whether related to sound, touch (heat, texture, pain), sight (brightness, color), and/or taste (maybe related to texture). This study demonstrates sensory sensitivities empirically — though with a small sample size — as in, in a lab setting versus general clinical evidence.
I think this is evidence of a fundamentally different way of processing sensory input — and therefore a different way of experiencing, seeing, and being in the world.
Sidebar: There is long debate about “identity first” versus “person first” language — simply put, is “person with autism” or “autistic person” more PC? Most of the autistic people I know personally prefer “autistic” and so that’s what I generally use in writing. See Jim Sinclair’s 1993 article Don’t Mourn for Us.
4. We can do better at providing education to autistic students
Primary education in the U.S. is a public good. And we have also passed laws providing pretty meaningful incremental money — mostly IDEA funding, but some others — towards students with disabilities.
The classic classroom environment isn’t great for everyone — or maybe anyone, when AI is upending venerable institutions like homework and conformity is rewarded much less than before — but it’s definitely not great for someone with a lot of sensory sensitivity.
For many autistic students, the delta between what they are capable of and what the structures around them allow for is huge. We can do so much better.
We’re also leaving a lot on the table for everyone. For every autistic employee at Goldman Sachs … there’s 10 or 100 autistic kids who could have been there too.
Technologies like VR and AI have the potential to support autistic students — as for all students, but some of the autistim-specific applications might include moderating strong stimuli, such as through moderating strong stimuli, practicing communication and social skills, or providing alternatives to behavioral therapy.
5. Building in this space is a good enough idea to get some funding
There’s a large and growing TAM due to increasing diagnosis prevalence.
There’s a stable government funding pool, both in education (e.g., IDEA) and in healthcare (e.g., Medicaid).
The competing services can be pretty poor. Check out Google Reviews of ABA (Applied Behavioral Analysis) providers near you.
PE firms are buying up the better ABA providers due to the TAM and public funding.
There’s white space in the use of technology because of the gap between what’s possible and what’s mainstream.
6. See also
Some startups / VC working in autism:
Some books:
- In a Different Key
- Those They Called Idiots: The Idea of the Disabled Mind from 1700 to the Present Day
- The Reason I Jump
- The Journal of Best Practices
- Funny, You Don’t Look Autistic
To add:
- How autism became the latest target for disruption
- This is also a great reference for some of the funding statements that I made w/o a source to cite
- Also lists some cool startups that I didn’t know about