INTERNATIONAL FINAL 2026–27 / ZHUHAI, CHINA / 23–26 JANUARY 2027

A.I. PROJECT

Project Themes & Requirements

From teaching a robot to building responsible AI solutions. Find your grade group below and explore your scenario, task, deliverables and criteria to win.

Grade K

Grades 1–3

Grades 4–6

Grades 7–9

Grades 10–12

01 / GRADE K

Visual Classification & Feature Recognition

Teaching a Robot to Sort My Room

Core AI focus: Computer Vision concepts, Data Labeling, and Feature Extraction.

The scenario

Students help a fictional “cleaning robot” learn the difference between “toys” (to be saved) and “trash” (to be discarded).

Your task

Students draw or cut out 5 images of toys and 5 images of trash from magazines to create a physical “training dataset.”
Acting as human “Data Labelers,” they must tag the images using stickers (e.g., green stickers for toys, red stickers for trash).

WHAT TO SUBMIT

A photograph of the sorted training data and a 30-second audio/video clip explaining: “I taught the robot that blocks are toys because they are colorful and stackable, while crumpled paper is trash because it can’t be played with.”

Criteria to win

Consistency in defining attributes/features for classification, demonstrating a fundamental grasp of how training data dictates machine learning accuracy.

02 / GRADES 1–3

Algorithmic Logic & Conditionals

The Self-Driving Maze Bot

Core AI focus: Sequential Logic, Rule-Based Systems, and If-Then-Else Conditional Branches.

The scenario

Students must design a no-code conditional logic algorithm to guide an autonomous delivery vehicle safely through a grid-based maze filled with obstacles.

Your task

Using a physical or digital grid map provided by the committee, students arrange directional and conditional command blocks (e.g., IF obstacle ahead → THEN turn right; ELSE → move forward 1 step).

WHAT TO SUBMIT

A clear visual flowchart or sequence map of their command chain, paired with a 1-minute video demonstrating how their algorithm handles sudden roadblocks or dead ends without crashing.

Criteria to win

Logical rigor and fault tolerance—the algorithm must successfully navigate the map without getting stuck in infinite loops or causing a collision.

03 / GRADES 4–6

Prompt Engineering & Generative AI Evaluation

The Endangered Species Advocate

Core AI focus: Structured Prompting, Iterative Refinement, and Identifying AI Hallucinations.

The scenario

Students act as environmental campaigns managers using text and image Generative AI tools to design an awareness poster and storyline for a local endangered animal.

Your task

Write structured prompts (defining role, context, task, and constraints) to generate educational copy and matching wildlife graphics.
Document a “Prompt Evolution Log,” highlighting where the AI failed early on (e.g., visual hallucinations like extra limbs, or factual text errors) and how they adjusted their prompts to correct those errors.

WHAT TO SUBMIT

The final generated awareness poster along with a detailed multi-round Prompt Iteration Log (minimum of 3 documented rounds of refinement).

Criteria to win

Active avoidance of raw, first-round AI outputs. Winners are selected based on their critical evaluation of AI limitations and their mastery over precise prompt parameters.

04 / GRADES 7–9

Predictive Modeling & Data Science

The Smart Campus Energy Predictor

Core AI focus: Tabular Data Analysis, Feature Engineering, Model Training, and Error Evaluation.

The scenario

A school wants to lower its carbon footprint and power bills. Students are given a historical 2-year virtual dataset containing variables like daily outdoor temperature, rainfall, campus attendance numbers, day of the week, and total daily kilowatt consumption.

Your task

Using low-code/no-code machine learning platforms (such as Orange, Teachable Machine, or entry-level Python libraries), students build a regression model to predict future energy demands. They must identify which features carry the most weight and test their model against a validation set.

WHAT TO SUBMIT

A 4-page data science report breaking down their data preprocessing steps, model accuracy metrics (e.g., Mean Absolute Error), and a data-backed recommendation brief for the school board’s energy policy.

Criteria to win

Analytical depth. High-tier scoring requires identifying data anomalies/outliers, preventing overfitting, and offering a robust economic explanation of feature correlations.

05 / GRADES 10–12

Applied Edge AI Solutions & Ethics Briefing

AI Assistive Ecosystems for Vulnerable Populations

Core AI focus: Deep Learning application (CNN/NLP/LLM API Integration), System Prototyping, and AI Safety, Bias & Privacy.

The scenario

Students must design and build a working AI application prototype tailored to solve a verified daily obstacle faced by vulnerable demographics (e.g., the visually impaired, isolated elderly, or non-verbal children).

Your task

Technical Implementation: Build a functional software prototype using open-source models or developer APIs (e.g., Hugging Face, OpenCV, Mediapipe). Examples include computer-vision navigation via edge cameras, or real-time speech-to-sign conversion.
Ethical Framework: Write a standalone comprehensive audit assessing dataset bias, user data privacy protocols (e.g., on-device processing vs. cloud exposure), and liability mitigation plans for system failure.

WHAT TO SUBMIT

A repository link (e.g., GitHub) containing clean, functional code alongside a 3-minute video demonstration of the prototype, accompanied by a 1,500-word Technical & AI Governance Whitepaper.

Criteria to win

The seamless integration of technical execution with deep humanitarian consideration. The system must actually execute its core loops, and the accompanying whitepaper must show mature, critical insight into deployment risks and regulatory compliance.

Questions about your A.I. project?

Contact the WOO! organizing team on WhatsApp

WOO! International Final 2026–27 · Zhuhai, China · 23–26 January 2027