AI for Youth Academy Future Scholars Research Initiative

Summer Camp · Coming Summer 2026

AI Explorer · Eight Sessions of Curiosity and Creation

We are finalizing the first-year AI Explorer summer experience for middle school innovators. Each session blends storytelling, guided experiments, and reflection so learners build confidence with intelligent tools before diving into advanced pathways.

Session-by-session preview

The AI Explorer camp unfolds across eight 3-hour labs. Facilitators can adapt timing for day or residential formats, but each session keeps the same mix of mini-lessons, studio time, and reflection.

  1. 01

    What Is Intelligence?

    Biological neurons vs. neural networks

    Key concepts

    • Differentiate natural and artificial intelligence
    • Structure of biological neurons and perceptrons
    • How signals and weights influence decision making

    Hands-on build

    • Build a paper neuron model that routes signals with weights
    • Role-play a mini neural network to classify simple inputs

    Reflection & share-out

    • Write a quick comparison of how brains and machines learn
    • Capture one question about intelligence for future exploration
  2. 02

    AI Around Us

    Weak vs. strong AI and the Turing Test

    Key concepts

    • Spot everyday narrow AI systems and their goals
    • Debate strong AI aspirations and limitations
    • Understand the Turing Test and alternative evaluations

    Hands-on build

    • Run a “guess the bot” chat activity in pairs
    • Map community touchpoints where AI already appears

    Reflection & share-out

    • Discuss where human judgment should stay in the loop
    • Exit ticket: one benefit and one risk of pervasive AI
  3. 03

    Teachable Machine: Vision

    Image classification project

    Key concepts

    • Training vs. testing examples for image models
    • Recognizing overfitting through live demos
    • Iterating on labels to improve clarity

    Hands-on build

    • Collect a quick image dataset with props and drawings
    • Train and export a Teachable Machine image model

    Reflection & share-out

    • Analyze misclassifications and suggest improvements
    • Document a short narrative of the project in a portfolio
  4. 04

    Teachable Machine: Sound & Pose

    Audio cues and movement recognition

    Key concepts

    • How sensors capture audio and skeletal data
    • Comparing model performance across modalities
    • Designing signals that are inclusive and dependable

    Hands-on build

    • Train separate Teachable Machine audio and pose models
    • Merge both models into a mini interactive experience

    Reflection & share-out

    • Evaluate which input type fit the challenge best and why
    • Plan a refinement list based on tester feedback
  5. 05

    Data Quality Matters

    Garbage in, garbage out

    Key concepts

    • Qualities of reliable datasets: balance, variance, relevance
    • Detecting bias and gaps before training
    • Ethical data sourcing and consent basics

    Hands-on build

    • Audit sample datasets for bias using a structured checklist
    • Run quick experiments showing how corrupted data hurts models

    Reflection & share-out

    • Create a team data quality charter for future projects
    • Journal prompt: how would poor data impact your community?
  6. 06

    Intro to Large Language Models

    Prompts, limits, and hallucinations

    Key concepts

    • Token patterns and why LLMs sometimes hallucinate
    • Prompt engineering basics: context, instructions, examples
    • Responsible use guidelines for generative text tools

    Hands-on build

    • Compare outputs from varied prompts in a shared doc
    • Design a fact-check protocol to catch hallucinations

    Reflection & share-out

    • Summarize a best-practice prompt formula in own words
    • Capture a case where human review is essential
  7. 07

    Scratch Lab: Linear Regression

    Visualizing predictions with blocks

    Key concepts

    • Representing datasets in Scratch lists
    • Calculating trendlines to make predictions
    • Interpreting slope and intercept in context

    Hands-on build

    • Code a Scratch project that fits a line to student-created data
    • Showcase predictions and error visually on stage

    Reflection & share-out

    • Explain how changing data shifts the regression line
    • Peer feedback on clarity of the Scratch project demo
  8. 08

    Scratch Studio & Ethics Forum

    Extend builds and debate AI responsibility

    Key concepts

    • Integrating sensors, APIs, or datasets into Scratch experiments
    • Ethical frameworks: fairness, accountability, transparency
    • Project storytelling and audience engagement

    Hands-on build

    • Iterate on Scratch AI projects with new features or datasets
    • Host an ethics circle using real-world case studies

    Reflection & share-out

    • Prepare showcase artifacts (demo video, poster, or write-up)
    • Personal pledge on how to build responsible AI in the future