Developing AI Agents for Disability and Self Determination: Lessons from Project RAISE
summary
Project RAISE builds a narrow agent called ZB that students select in a competition and help train by recording third‑grade language; facial tracking is abandoned for an ethical halo and biometric heart‑rate input lets ZB modulate cadence. Three phases—learn to code, teach a peer, then classroom—run across five years and a 150‑student randomized study; measured gains include communication, on‑task behavior and math. Design rule: built with not for, treating agents as a cognitive prosthetic.
AI-written summary · about
chapters (50)
- 0:00Developing AI Agents for Disability and Self-Determination
- 1:33Built For vs Built With - The Word That Changes Everything
- 2:35Speaker Intro - Dr. Lisa Dieker, Williamson Family Distinguished Professor in Special Education at University of Kansas
- 3:38Why a Special Education Professor Is Deep in Technology
- 4:39Life Lessons: A Son With Tourette's, Dyslexia and Dysgraphia
- 5:09Self-Determination Begins With Voice - Have You Asked Them?
- 5:39Bookshare, Kindle and What Actually Changed His Life
- 6:42The Mistake: Retrofitting Instead of Co-Designing
- 7:12AI as a Cognitive Prosthetic, Not a Replacement for Learning
- 7:44What 24 PhD Students With Disabilities Taught Her About Accessibility
- 8:14Not Every Disability Is Visible
- 8:45The Accessibility Paradox: AI Helps and Complex Interfaces Don't
- 9:15Prompts Written Above a Fifth-Grade Reading Level
- 9:45Uneven Representation and the Wheelchair Cliché
- 10:18The Data: 16% Disability Representation in AI Design
- 10:48What's Missing: Reliable Multimodal Access
- 11:50Cognitive Access and Translating Content Down, Not Just Across
- 12:21Trustworthy Human Judgement and When an Agent Should Not Decide
- 12:52Who's Missing: The Design Team and the Data Sets
- 13:23Why Project RAISE Stopped Using Facial Tracking
- 13:54Designing With Disability First Makes It Better for Everyone
- 14:24The Microwave, the Car, the Phone, the Computer - Design Assumptions
- 15:55Shifting From Build-Then-Fix to Co-Design From the Start
- 16:59Flipping the Model: Evaluate and Improve Before Market
- 17:30TeachLive: The First Avatar Simulator for Training Teachers
- 18:01Martin and Bailey - Avatars Co-Designed With Real Families
- 19:02Project RAISE: Teaching Kids With Autism to Code a Robot
- 20:02The First Agent the Kids Said Looked Stupid
- 20:34Running a Competition So the Students Chose the Agent
- 21:07Recording Third Graders So the Agent Used Their Actual Language
- 21:39Replacing Facial Tracking With a Halo on Ethical Grounds
- 22:41ZB and RAISY: A Narrow Agent Built for a Specific Need
- 23:12The Three Phases: Learn, Teach a Peer, Then Into the Classroom
- 24:13Adding Biometrics So the Agent Responds to a Rising Heart Rate
- 25:15Affirmation Over Praise: "You Got This" Instead of "Great Job"
- 25:45Results: Communication, On-Task Behaviour and Maths Gains
- 26:16Where AI Agents Could Go: Goal Setting, Self-Advocacy, Transition
- 26:47Support Decisions - Don't Make Them for People
- 28:18Current Work: Coaching Teachers With Multimodal Biometric Data
- 29:51FLITE ICA: Fully Automated Audio Analysis in Under Two Minutes
- 30:54Toward the Neurophysiological State of Teacher Mind and Body
- 31:54What She's Learned: Compliance Standards Are Not Accessibility
- 32:26The Future Isn't Artificial Intelligence - It's Amplified Self-Determination
- 33:28Q&A: Securing an Agent That Acts Like an Employee
- 35:00Q&A: Supporting Independence Without Creating Dependency
- 36:01Q&A: Which Disabilities Are Hardest to Design For
- 37:33Q&A: Accountability for an Autonomous Agent's Security Incident
- 39:06Q&A: Ethical Lines - Representation, Guardrails and Closed Loops
- 40:07Q&A: On "Challenged" vs "Disability"
- 41:09Q&A: Designing for Senior Citizens and Ageing Users
related talks
21mMeet your new and improved teammate! Human-directed AI agents and gaming
51mBuilding Responsible Autonomous AI Agents: Principles, Patterns, and Practices
28mAgentic AI
29mPairing with the machine: How AI Agents are rewriting the rules of software development
27mHow to Build AI Agents and Keep Your Sanity
26mAI Agents and the Future of Human-AI Collaborative Intelligence
52mRead, Write, and Tinker with AI
29mBuilding Trust in AI Agents