A human-centric AI experience where people learn to build, train, challenge, and govern their own AI Reps.
Developed over a decade with technologists, educators, students, designers, and civic innovators.
01 — The opening problem
AI is already changing how people are seen, sorted, supported, evaluated, and represented. Yet most AI tools ask people to adapt to the system. They do not give people meaningful ways to train the AI, question it, correct it, or decide what it should remember.
People should not just use AI. They should help shape the AI that represents them. AI that belongs to them.
That means AI literacy through practice, not lectures. Data dignity made real, through experiential learning. Tools to curate your own digital memory, set boundaries, and test how your AI understands you.
02 — What it is
Gameshow is a concept-stage, prototype-informed platform for building AI Reps: personal, human-guided AI representatives that learn from a person's own stories, choices, reflections, and feedback. Users don't simply chat with AI. They train it. They challenge it. They correct it. They decide what it can remember.
Voice
Expressing who you are and what matters to you.
Agency
Shaping how AI learns from you and acts with you.
Dignity
Controlling what becomes memory, what stays private, and what may represent you.
03 — The AI Rep
Not a generic assistant. A personal digital representative trained through your approved stories, reflections, goals, boundaries, and corrections.
A Rep should not guess who you are from hidden data. It should learn through an accountable loop.
Step 01
A person speaks.
Step 02
The system reflects.
Step 03
The person corrects.
Step 04
The person decides what becomes memory.
Step 05
The Rep practices.
Step 06
The person gives feedback.
Step 07
The Rep improves.
Not a profile.
A Rep is built from approved memory, not hidden surveillance.
Not a chatbot.
A Rep is trained through feedback, not just conversation.
Not an avatar.
A Rep is about agency, memory, and representation — not appearance.
Not autonomous by default.
A Rep earns trust through correction, review, and permission.
04 — How it works
Casting Call finds the voice. Backstage governs the memory. The Arena trains the representation.
Find your voice.
Voice-first prompts, interviews, stories, and reflections. Not a personality quiz, not a profile screen — a space for discovery. What do I care about? How do I explain myself? What should my Rep understand, and what should it be careful about?
Govern the memory.
Where stories become draft memory. Keep, edit, delete, make private, share, expire, or turn a reflection into a practice challenge. Data dignity made visible — you are not generating data, you are editing and governing the memory your Rep is allowed to use.
Test the representation.
Where the Rep practices. Rehearse a presentation, explain a project, answer questions, respond to a challenge. You see what it gets right, what it gets wrong, and how your feedback changes the next attempt. Private first — user and Rep — before any social mode.
05 — What makes it different
The DOTES framework — Do · Observe · Tell · Explore · Show — connects early narrative forms of learning with later scientific practices. A developmental continuum from intuitive storytelling to formal inquiry, used here to turn lived experience into structured, reviewable memory.
AI literacy through doing
Learn how AI responds to training, feedback, memory, and constraints by experiencing it directly — not by reading about it.
Personal AI Reps
Each Rep is shaped by a person's own voice, stories, goals, and boundaries. Not generic personalization — accountable representation.
DOTES
Do, Observe, Tell, Explore, Show. A framework that turns lived experience into structured, reviewable memory instead of raw chat logs or behavioral exhaust.
Data Backpacks
A user-controlled container for approved memories, reflections, artifacts, permissions, and Rep state. Carry your meaning forward — don't get locked inside a platform.
Human feedback as the engine
The Rep improves because people correct it. Correction is not a failure state. It is the most important learning signal in the system.
Privacy-first by design
Raw reflection is private by default. Durable memory requires review. Representation requires permission. You can always see and change what the Rep thinks it knows.
06 — Why it matters now
For people
A way to build confidence, voice, reflection, and agency in an AI-shaped world.
For educators and mentors
A practical experience for teaching AI literacy, digital identity, storytelling, feedback, and responsible technology use.
For builders and researchers
A living testbed for human-centered AI, consent-based memory, personal knowledge graphs, and safer AI representation.
07 — Where the project is now
Gameshow is still early — and that is the opportunity. The foundations are in place: AI Reps, DOTES, Data Backpacks, the three-space experience, human feedback loops, and a responsible AI posture grounded in consent, privacy, safety, and agency. Now the work is to build the first trustworthy version.
This is not a finished platform. It is a build invitation — a chance to help shape the product, governance, learning model, safety architecture, and first real user experience before the system hardens.
08 — The invitation
The question is not whether AI will know more about us. The question is whether we will have a say in how AI represents us. If that question matters to you — as a person, an educator, a builder, a researcher — there is a place for you in this work.