Research map / Large language models / AI and society
AI Chef Trainer: Introducing Students to the Importance of Data in Machine Learning
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Presents AI Chef Trainer, a Flask web app where middle schoolers test a Naive Bayes recipe recommender, add their own recipes, retrain it, and retest to learn about the role of data in ML.
Problem
Children need AI literacy, specifically an understanding of how training data affects ML system behavior, and few hands-on tools make this concrete for K-12 students.
Approach
A cyclical test-contribute-retrain-retest educational app built around recipe recommendation, where students' own ingredient-recipe pairs are added to a deliberately small dataset and used to retrain a TF-IDF plus Naive Bayes classifier, making data quantity and specificity observable.
Key results
- 45 of 52 students (87%) submitted at least one recipe, and 28 submitted two or more.
- 26 of 52 students (50%) tested the AI with subsets of their training ingredients or full ingredients from their own recipes.
- Students' qualitative responses (e.g., 23 on data quantity, 11 on data specificity) showed recognition that more and more specific ingredients improve recommendations.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use AI Chef Trainer for recipe testing and recipe contribution | Middle school students · Middle school study | students submitting at least one recipe: 45 of 52 (87%) | Helped |
| Use AI Chef Trainer for recipe testing and recipe contribution | Middle school students · Middle school study | students testing with their own recipes: 26 of 52 (50%) | Helped |
| Use AI Chef Trainer for recipe testing, contribution, and retraining | Middle school students · Middle school study | understanding of data quantity and specificity: 23 responses; 11 students | Helped |
| Compare outcomes across grade levels | AI Chef Trainer · Middle school study | student outcomes | No clear effect |
Surprising findings
- Students were amused and tested unusual combinations deliberately, e.g. a student's ice cream recipe with meat and beef was accepted as the top recommendation after retraining.
- The software did not check for repeated ingredients or case-sensitivity, which affected recommendations.
Limitations
- Students interacted for only 8-10 minutes, limiting depth of learning.
- Study sample was a single STEM charter school, and 11 of 63 students were excluded from analysis.
- Evaluation is descriptive, based on logs, surveys, and qualitative quotes, without a controlled comparison.
Experimental setup
Datasets: Deliberately limited base dataset of 79 recipes Student-contributed recipes from middle school study
Metrics: student engagement counts (recipes submitted, tests conducte survey responses ingredient counts in post-training tests
Builds on: Google Teachable Machine Machine Learning for Kids ChemAIstry
Closest papers by meaning
- An XAI Social Media Platform for Teaching K-12 Students AI-Driven Profiling, Clustering, and Engagement-Based Recommending (AAAI 2025 · similarity 0.82)
- A Toolbox for Modelling Engagement with Educational Videos (AAAI 2024 · similarity 0.81)
- “AlphAI”: Teaching AI Algorithms to K12 by Training Learning Robots and Visualizing How It Works (AAAI 2025 · similarity 0.80)
- Smart Motor: A Low-Cost Hardware and Software Toolkit for Introducing Supervised Machine Learning to Elementary School Students (AAAI 2025 · similarity 0.80)
- DataEnvGym: Data Generation Agents in Teacher Environments with Student Feedback (ICLR 2025 · similarity 0.80)
- Towards an AI Course Based on Neural Networks (AAAI 2025 · similarity 0.80)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.