Problems → Prototypes · AILE
A Curiosity-Preserving AI Mentor
Personalized AI mentors may optimize completion, hints, and immediate success while unintentionally reducing exploration, learner agency, or tolerance for difficulty. A system that predicts the learner too aggressively can become a path optimizer instead of a partner in discovery.
This exercise runs locally using simple rules. Your entries stay on this page and clear when you reload. No AI service is called.
Critical review of this working prototype
Strongest assumption: Authored challenge choices plus attempt/reflection cycles support exploration without taking over learner decisions.
Likely failure: A journal rewards completion or lets learners switch away from every difficulty without explaining evidence; a revised attempt can inherit stale reflection.
Test that could change the design: Compare with a paper choice journal on a new project. Reject if cycle completion rises while independent explanations, willingness to test ideas, or transfer decline.
Non-AI comparison: A human mentor and a learner-owned paper experiment journal with challenge cards.
Evidence to collect: Human review of prediction/test quality, evidence-based reflection, later independent project complexity, and voluntary path-change reasons. Cycle count is not a learning score.
Human control and access: Authored paths center game design; the own-challenge option allows other interests. No ability inference or automatic simplification. Text descriptions of oral, visual, and coding attempts are allowed; no upload, telemetry, or automatic sharing.
Changes made during review
- Required an actual recorded attempt before exposing optional hints and reflection.
- Required a reflection and reason for the learner-selected next step.
- Preserved selected path even when the learner chooses switching, pausing, or human support.
- Invalidated active cycles on goal/path edits and cleared stale reflections when attempts change.
- Retained completed cycles in a local exportable journal and cleared them on reset.
- Disabled no-JavaScript input/submission and completed all eight design-cycle steps.
- Smallest next experiment: a mentor reviews one attempt and an unseen transfer task with the learner.
This is a design and implementation review, not an empirical validation of learning outcomes.
Take the brief into your own AI environment
Copy a self-contained review prompt, then open the AI workspace you already use. The prompt asks for a falsification test, a non-AI alternative, evidence that distinguishes activity from learning, and human-system risks.
Preview the prompt
Act as a critical learning-engineering reviewer. Review the prototype brief below as a thin-slice experiment, not as a finished product. Return: 1. The strongest design assumption. 2. The most important plausible failure mode. 3. The highest-value falsification test for the next cycle. 4. One credible non-AI alternative that could address the same capability gap. 5. The next smallest prototype worth building. 6. What should be measured to distinguish engagement from learning, transfer, and real performance. 7. Human-agency, equity, accessibility, privacy, or governance concerns that should change the design. Be concrete. Separate evidence-backed claims from hypotheses. Do not reward novelty for its own sake, and do not assume more AI is better. COLLECTION: AILE TITLE: A Curiosity-Preserving AI Mentor CAPABILITY GAP: Personalized AI mentors may optimize completion, hints, and immediate success while unintentionally reducing exploration, learner agency, or tolerance for difficulty. A system that predicts the learner too aggressively can become a path optimizer instead of a partner in discovery. FIRST PROTOTYPE: Design an AI mentor that begins with learner-chosen goals, uses short challenge cycles, offers optional creative pathways, and periodically asks the learner to decide what to pursue next. Human mentors can see the rationale for interventions and override the system. REPRESENTATIVE USE CASE: A teenager learning game design alternates between coding challenges and visual design choices. The AI notices repeated debugging difficulty but offers two next-step paths rather than automatically simplifying the task. LEARNING / HUMAN-SYSTEM FRAME: The design supports autonomy, competence, metacognition, and human oversight. Personalization is treated as a hypothesis about support, not a hidden judgment about ability or motivation. QUESTIONS ALREADY IDENTIFIED: 1. How do we measure curiosity without reducing it to clicks or facial signals? 2. When should the mentor challenge rather than accommodate? 3. Does learner choice improve persistence and transfer over time? FIRST-CYCLE FRAMING: Understand: Validate that the real gap is the ability to pursue increasingly complex learning with agency, reflection, and sustained curiosity, not simply low engagement, low tool use, or a workflow inconvenience. Map: Map the system: learner, human mentor, chosen project, AI mentor, peer/community context, and progression model. Identify where the capability currently succeeds, breaks down, or is masked by other constraints. Instrument: Predefine evidence: independent project complexity, persistence, self-explanation, choice patterns, transfer, and mentor override. A key disconfirming signal is: the mentor keeps learners active but narrows exploration or makes challenge avoidance easier.
Open the underlying design brief and eight-step path
A first prototype
Design an AI mentor that begins with learner-chosen goals, uses short challenge cycles, offers optional creative pathways, and periodically asks the learner to decide what to pursue next. Human mentors can see the rationale for interventions and override the system.
Representative use case
A teenager learning game design alternates between coding challenges and visual design choices. The AI notices repeated debugging difficulty but offers two next-step paths rather than automatically simplifying the task.
Questions worth carrying forward
- How do we measure curiosity without reducing it to clicks or facial signals?
- When should the mentor challenge rather than accommodate?
- Does learner choice improve persistence and transfer over time?
Evidence anchors
These sources motivate the design; they do not validate this prototype.
- Niemiec, C. P., & Ryan, R. M. Autonomy, competence, and relatedness in the classroom: Applying self-determination theory to educational practice. (2009)
Motivation, learner agency, and challenge/support balance.
- UNESCO. AI Competency Framework for Students. (2024)
Student AI literacy, critical judgment, and co-creation.
- Amershi, S., et al. Guidelines for Human-AI Interaction. CHI 2019. (2019)
Human-AI interaction design and calibrated reliance.
- Education Endowment Foundation. Metacognition and Self-Regulated Learning, 2nd ed. (2025)
Metacognitive prompts, self-monitoring, and fading support.
- Chi, M. T. H., & Wylie, R. The ICAP Framework: Linking Cognitive Engagement to Active Learning Outcomes. (2014)
Designing learner actions rather than optimizing superficial engagement.
One possible first pass
- 01 Understand
Validate that the real gap is the ability to pursue increasingly complex learning with agency, reflection, and sustained curiosity, not simply low engagement, low tool use, or a workflow inconvenience.
- 02 Map
- 03 Design
Compare an AI intervention with simpler non-AI options. Specify what the human decides, what the AI may suggest, and which tradeoffs are acceptable.
- 04 Build
Create the smallest usable prototype around one representative task, with expert-curated content, visible uncertainty, and an easy human override.
- 05 Instrument
Predefine evidence: independent project complexity, persistence, self-explanation, choice patterns, transfer, and mentor override. A key disconfirming signal is: the mentor keeps learners active but narrows exploration or makes challenge avoidance easier.
- 06 Deploy
Pilot with a small, representative group in a near-real setting; preserve a baseline or comparison condition and document implementation conditions.
- 07 Evaluate
Look for capability growth and transfer, not just satisfaction or activity. Inspect subgroup patterns, human workload, errors, and unintended adaptations.
- 08 Refine
Let the evidence change the problem model. Default back to UNDERSTAND if the assumed gap was wrong; otherwise revisit the earliest step invalidated by the evidence.