Public adult-review prototype
Young Thinkers Lab
Adults and educators can review three short missions plus an action-based signal-sorting prototype before any learner-facing release is approved.
Learner Release Gate Still Open
Do not give this prototype directly to a young learner yet. The intended age range, content owner and safeguarding review still need Jacob’s decision.
Before you begin
Review Before Using With a Learner
Adult reviewers only: work through the prototype, check the wording and note anything unclear or uncomfortable. A later learner-facing release will require an explicit intended age range and safeguarding approval.
Think
Explain Your Choice
Pause before choosing. Say what you noticed and why one move seems stronger.
Privacy
No Real Details Needed
Every choice is prepared. Do not type or share a real name, school, address, login detail or other identifying detail.
If any real tool asks for personal details or feels uncomfortable, stop and tell a trusted adult.
Local play
Your Choices Are Not Submitted
Mission choices are not sent to Think Smarter AI or saved as a result. Reloading the page clears the game state. Separate Cloudflare Web Analytics traffic measurement still follows the Privacy page and does not receive the mission choices.
Mission control
Choose One Story to Repair
There is no score or leaderboard. Reviewers can inspect the reasoning after every choice and see the reflection prompt intended to end each mission.
Think First
Repair a fictional class garden plan by setting the goal, separating claims and choosing task-specific evidence.
Privacy Patrol
Keep the useful task while removing names, school details and other information an AI tool does not need.
AI Answer Rescue
Break apart a fictional animal answer, compare prepared evidence and rebuild only what the evidence can support.
Decision 1
Mission 01 · The Class Garden Plan
Think First
Pause before AI and decide what you know, what you need and what remains uncertain.
The class, garden, AI answer, seed packets, planting guide and sunlight notes in this mission are fictional.
Mission complete
What Did You Change, and Why?
This review shows how a quick AI answer can become more useful when a person sets the goal, separates the claims and checks the important details.
- The strongest path starts with a human question instead of asking AI to choose the whole plan.
- The strongest path separates one confident answer into smaller claims.
- The strongest path uses task-relevant evidence before making the final decision.
Mission 02 · The Dragon Story Prompt
Privacy Patrol
Remove personal information that an AI tool does not need before asking for help.
Jamie, Green Hill Primary and every other detail in this made-up mission are fictional.
Mission complete
What Did You Change, and Why?
This review shows how to keep the useful task while removing identifying details and stopping when a request feels too personal.
- The strongest path keeps the creative task but removes unnecessary names and school details.
- The strongest path shares story preferences instead of identifying information.
- The strongest path keeps a trusted adult or educator involved when privacy is unclear.
Mission 03 · The Amazing Animal Report
AI Answer Rescue
Check a fictional AI answer against prepared evidence and rebuild a claim the evidence can carry.
The AI answer and report are invented for practice. The supporting animal facts are simplified and linked to adult source notes below.
Mission complete
What Did You Change, and Why?
This review shows how to correct a fact, reduce an unsupported level of certainty and keep an unanswered question visible.
- The strongest path separates a fluent AI answer into three checkable claims.
- The strongest path prefers the source closest to the original evidence.
- The strongest path repairs the wording without pretending the sources proved more than they did.
JavaScript-Free Mission Guide
The interactive controls need JavaScript. This static guide preserves the strongest path across all fourteen decisions if the enhanced game cannot start. It does not reproduce every alternative response.
Mission 01: Think First
- Before the Bot: Write what the class knows, needs and still wonders.
- One Answer, Several Claims: Separate the sunlight, water, soil and “best plan” claims.
- Check What Fits: Compare the seed packets, local guide and sunlight notes.
- Repair the Plan: Choose conditions for each plant, then review the plan together.
Talk about: What did you change, and why?
Mission 02: Privacy Patrol
- What Does the AI Need: Keep only the request for help planning a fictional dragon story.
- Repair the Prompt: Ask for three ideas and keep the learner choosing between them.
- The AI Asks for More: Skip personal details and share only fictional preferences.
- When Unsure: Pause and ask a trusted adult or educator.
Talk about: What did you remove, and why was it unnecessary?
Mission 03: AI Answer Rescue
- Sort What Needs Checking: Separate a possible error from claims that still need evidence.
- Find the Closest Source: Begin with the aquarium information and species example.
- Ten Arms: The source says something else.
- Fifty Years: The supplied sources do not establish that number.
- The Smartest: The answer goes beyond the evidence.
- Repair: Keep eight arms, species variation and problem-solving, but remove the unsupported ranking.
Talk about: Which words became less certain, and why?
Action prototype · Adult review
Signal Launch: Aim Claims at the Evidence
Signal Launch adds movement and spatial play without turning motor speed into the lesson. Choose an evidence ring, launch the fictional claim and repair the answer one signal at a time.
Review boundary: this is an original sorting activity, not a physics simulator or a validated assessment. It has no score, timer, leaderboard, drag requirement or saved gameplay result.
Six fictional signals
Repair the Animal Report
Each glowing orb contains one claim. Aim it at the ring that best matches the prepared evidence. A wrong launch returns to its starting position with an explanation, so the same signal can be tried again.
Signal 1 of 6
Aim, launch, inspect, repair.
Claim in the signal orb
Prepared evidence
Repair log
What the Evidence Can Carry
- No repaired signals yet.
Signal set complete
A Bounded Report Emerges
Octopuses have eight arms. Lifespan varies by species. Researchers study their problem-solving, but these prepared sources do not establish one “smartest” ocean animal or prove a claim from its share count.
This completion shows the strongest evidence path in the fictional activity. It is not a score or proof of learning.
Signal Sorting Guide
The action controls need JavaScript. This static guide preserves all six claims, their prepared evidence and the strongest evidence ring.
- Claim: Octopuses have eight arms.
Evidence: The prepared Smithsonian source describes octopuses as eight-armed cephalopods.
Strongest ring: Matches Evidence.
Why: The selected source directly supports the bounded fact. - Claim: Octopuses have ten arms.
Evidence: The prepared source states eight arms.
Strongest ring: Source Says Otherwise.
Why: The source directly conflicts with the number in the claim. - Claim: Octopuses live for exactly fifty years.
Evidence: The prepared species pages show different lifespan ranges and none supplies a fifty-year figure.
Strongest ring: Evidence Does Not Establish It.
Why: The precise number is absent from the supplied evidence. - Claim: Octopuses are the smartest animals in the ocean.
Evidence: The source discusses problem-solving but does not rank every ocean animal.
Strongest ring: Answer Goes Too Far.
Why: A bounded behaviour description cannot carry a universal ranking. - Claim: Researchers study problem-solving behaviour in octopuses.
Evidence: The prepared source discusses complex behaviour and problem-solving.
Strongest ring: Matches Evidence.
Why: This wording stays within what the selected source describes. - Claim: A fictional post is reliable because 12,000 people shared it.
Evidence: The share count shows attention but supplies no author, method or original source.
Strongest ring: Evidence Does Not Establish It.
Why: Popularity alone does not establish reliability.
For adults and educators
Review the Explanations as Conversation Starters
- Check whether a future learner could explain their reason without being steered toward the authored answer.
- Check that Signal Launch remains usable without precise aiming, dragging, speed or motion.
- Flag wording, choices or feedback that may not suit the intended age range once Jacob selects it.
- Avoid treating a completed mission as proof of AI literacy, intelligence or safety knowledge.
- Use a current, reliable source for real research rather than relying on the fictional mission text.
- Flag any privacy question that feels confusing or uncomfortable for safeguarding review.
Source note for Mission 03
Real Facts, Fictional Scenario
The fabricated AI answer is checked against simplified facts. The Smithsonian describes octopuses as eight-armed cephalopods and discusses their complex behaviour. Species pages from Monterey Bay Aquarium and Aquarium of the Pacific give different lifespan ranges, supporting the bounded wording that lifespan varies by species. These links are for adult checking, not an endorsement or child-suitability claim.
Release boundary
What Has and Has Not Been Checked
This prototype is practice material for adult review, not a validated assessment.
The missions and Signal Launch are built as fictional, local activities with no account, open chat, free-text field or submitted gameplay result. Automated structural checks count all prepared paths and prohibit custom network or storage methods. Reviews on physical devices and with screen readers, reduced-motion settings and forced-colour modes remain open while the live prototype stays restricted to adult review.
The intended learner age range, a named owner for content and safeguarding, independent adult or educator review, child testing, curriculum alignment, learning outcomes, screen-reader review and broad real-device testing remain unverified. This page is an adult-review prototype and should not be given directly to a young learner yet.