The Inner Game of AI™ – Fluency & Navigation
A self-paced pathway for middle-school students to build AI skills: terminology, prompting, evaluation, safety, verification, and decision-making.
About the Program
The Inner Game of AI™ focuses on practical AI competency. Students learn core concepts, precise prompting, iterative improvement, verification, data awareness, and safe, effective use of AI tools.
Fluency (Pillars 1–10): Core skills—concepts, clarity, decomposition, iteration, experimentation, and error-driven improvement.
Navigation (Pillars 11–25): Decision skills—goal setting, source awareness, evaluation, trade-offs, safety, and human-in-the-loop workflows.
Each pillar includes a clear Definition, five Reflection questions, and a concise Closing statement.
Fluency Pillars (1–10)
“Fluency builds capability: understand the system, specify the task, iterate the method.”
Pillar 1 Flow of Knowledge
Definition
Understand AI as an input–process–output system. Learn basic terms (prompt, tokens, context window, model limits) and how information flows from your instructions to the model’s response. Treat AI like a programmable partner that follows patterns you describe.
Reflection
- What parts of a prompt directly control the model’s output?
- How would you explain “context window” in one sentence?
- Where does AI get patterns from, and what does that imply?
- What is the difference between instructions and data examples?
- When is a short prompt better than a long one?
“Know the system: inputs shape outputs.”
Pillar 2 Adaptability
Definition
Learn tool-agnostic skills you can transfer across apps and models. Switch formats (text, table, outline), adjust constraints (length, tone, audience), and port workflows between tools without losing quality.
Reflection
- How would you convert a paragraph answer into a table?
- What changes when you move a prompt from one tool to another?
- Which parts of a workflow are reusable anywhere?
- How do you keep constraints consistent across tools?
- What’s your plan if a feature is missing?
“Be tool-independent; keep the workflow portable.”
Pillar 3 Clarity
Definition
Specify tasks precisely: role, objective, constraints, format, and success criteria. Use examples and counter-examples. Remove ambiguity to reduce errors and variability.
Reflection
- What are the minimum fields of a clear task brief?
- How do examples change model behavior?
- When do you use bullet lists vs. numbered steps?
- What constraint most improves reliability?
- How do you define “done” for a task?
“Clear specs produce predictable results.”
Pillar 4 Applied Practice
Definition
Develop fluency through repetitions. Run prompt drills (summarize → outline → rewrite → check) and compare outputs. Track what changes improve accuracy or usefulness.
Reflection
- What drill can you repeat in five minutes a day?
- Which prompt edit produced the largest improvement?
- How will you log versions and outcomes?
- What metric will you use to judge quality?
- How many trials are enough before deciding?
“Skill grows from deliberate repetitions.”
Pillar 5 Creativity
Definition
Use structured ideation: SCAMPER, “many-then-filter,” and prompt branching. Generate multiple options, then score and select using criteria aligned to the task.
Reflection
- How would you branch one prompt into five variants?
- Which scoring criteria fit this assignment?
- When should divergence stop and selection start?
- How do constraints improve originality?
- What’s your method to combine top ideas?
“Create widely, then select with criteria.”
Pillar 6 Systems Connection
Definition
Map how pieces connect: data → prompt → model → output → verification. Recognize dependencies (e.g., poor data yields poor outputs) and plan handoffs between steps.
Reflection
- What is the weakest link in your pipeline?
- Where should verification occur?
- What inputs most affect output quality?
- How do you document handoffs between steps?
- Which step benefits most from automation?
“Know the pipeline; strengthen the links.”
Pillar 7 Inquiry
Definition
Form effective questions: define unknowns, assumptions, and hypotheses. Use iterative querying (broad → narrow) and ask for missing context explicitly.
Reflection
- What do you need to know first?
- Which assumption requires testing?
- How would you narrow a broad topic in three steps?
- What context should you force the model to request?
- How will you record answers for re-use?
“Good questions structure discovery.”
Pillar 8 Task Focus
Definition
Scope tasks precisely. Break large goals into atomic steps (extract → classify → transform → evaluate). Limit each prompt to one clear operation to reduce error.
Reflection
- How can you split this task into atomic steps?
- Which step should the model do first?
- What is the expected input and output for each step?
- Where can format templates prevent confusion?
- How will you detect task creep?
“Small, clear steps beat vague, giant ones.”
Pillar 9 Error-Driven Improvement
Definition
Treat mistakes as data. Classify failure types (hallucination, formatting miss, scope miss, source error) and apply fixes (cite requirement, schema, constraints, retrieval).
Reflection
- What failure type occurred here?
- Which constraint would prevent it next time?
- How can you surface uncertainty in the output?
- Where do you need external sources?
- What validation test will catch this earlier?
“Label errors, apply fixes, improve the loop.”
Pillar 10 Incremental Skill Building
Definition
Advance in small increments: version prompts, compare outputs, and keep a short “playbook” of working patterns. Reuse proven templates across tasks.
Reflection
- What is your version naming scheme?
- Which template delivers the most consistent output?
- How do you run A/B tests on prompts?
- What belongs in your personal playbook?
- How often will you review and update it?
“Version, compare, keep what works.”
For Schools
Free Access in Puerto Rico: Title I middle schools (Grades 6–8) in Puerto Rico receive the program at no cost.
Self-paced and 508-compliant—students can log in on any device.
Program value: $50 per student — fully donor-supported.
For Donors
Every contribution expands access to The Inner Game of AI™ for Puerto Rico Title I middle schools.
- $50 = Full program for 1 student (all 25 pillars).
- Secure processing · Tax-deductible under 508(c)(1)(A).
- Focused strictly on AI skills: fluency, evaluation, safety, and decision-making.
FAQ
Is this program free for schools?
Yes. Title I middle schools (Grades 6–8) in Puerto Rico receive free access for all enrolled students.
Who is eligible for free access?
Public/charter Title I middle schools in Puerto Rico. If you’re unsure, contact us at [email protected].
How do students access the program?
Schools apply; once approved, we provide a simple login link usable on school or home devices. Self-paced, mobile-friendly.
Is the program bilingual and 508-compliant?
Yes. English and Spanish (PR), built to follow Section 508 accessibility guidelines.
What’s included in the $50 value per student?
Full access to all 25 pillars (Fluency 1–10, Navigation 11–25) with definitions, reflections, and closings focused on AI skills, verification, and safe use.
How can donors support the mission?
Every $50 sponsors one student’s full program. Donations are processed securely and are tax-deductible under 508(c)(1)(A). Use the floating Donate button or this link.
“Fluency builds capability. Navigation directs application. Together they turn AI into a tool you can use with accuracy, safety, and judgment.”