Designing assistive AI with people, not for them
The principles guiding our exploratory AI for Autism research: agency, individuality, consent and humility about what technology can do.
CyberFoxX · 5 min read
Autism isn't a problem to fix. It's a language to understand. That idea shapes how we approach our AI for Autism initiative, which is exploratory research, not a product, a therapy or a medical device.
Principles we hold ourselves to
- Individuals first: no two autistic people have the same preferences or support needs.
- Agency and dignity: technology should adapt to the person, not demand that the person adapt to the technology.
- Consent: guardian authorisation, age-appropriate controls and safeguarding must be designed in before any wider use.
- Humility: AI cannot reliably infer someone's emotions from a camera, and we will not claim that it can.
- Privacy: sensitive information stays local and under the family's control.
Questions we want to explore
Could predictable, adjustable interfaces reduce friction in daily routines? Could a locally hosted companion help someone practise communication at their own pace? Could families set up routine support that respects the individual's choices? These are research questions, and they need neurodivergent people, families, clinicians and researchers in the room.
If you are a researcher, clinician, advocate or family interested in shaping this work responsibly, we would like to hear from you.