005. AIYA Impact Symposium 2025-2026

AIYA Journal

Aug 28

3 min read

The inaugural AIYA Impact Symposium showcased innovative student projects across the domains of healthcare, environmental science, infrastructure, education, and autonomous AI systems, spanning diverse applications that range from clinical measurement apps and automated industrial inspection systems to live environmental data platforms, educational workshops, health management tools, and multi-agent research frameworks. Representing schools across the U.S., Kuwait, and Australia, our featured projects share the common thread of successfully applying AI to solve real-world problems while tackling complex, open-ended challenges.

Featured Projects

WoundCorder

Contributors: Mitchell — Brentwood School
Overview: A smartphone app utilizing consumer LiDAR to measure wound volume with 90–95% accuracy across 60 test trials, overcoming traditional ruler-based limits. The project also benchmarked 19 vision-language models on wound characterization, uncovering systemic laterality (left/right) errors and testing prompt strategies to correct them. The work proves consumer hardware can deliver clinical volumetric data while exposing critical AI vision failure modes requiring human oversight.
Summary: Find project summary here
Video: Find project video here

Revos

Contributors: Lucas & Andrew — Brentwood School
Overview: An early-stage UAV inspection system designed to make oil well site monitoring faster, safer, and more consistent than manual reviews. Combining custom aircraft design, sensor planning, and AI-assisted computer vision, Revos automatically flags leaks and structural damage from aerial footage. Current development focuses on validating an end-to-end flight, detection, and human-verification pipeline for energy infrastructure operators.
Summary: Find project summary here
Video: Find project video here

CO2 Pomfret

Contributors: Dongheon — Pomfret School
Overview: A live, school-hosted web application that transforms a decade of hand-collected forest data (2015–2025) into actionable insights on tree growth and carbon sequestration. Built using AI acceleration via Cursor, the platform features dynamic trend dashboards and carbon-impact projection tools now actively integrated into Pomfret's forestry management operations and environmental curricula.
Summary: Find project summary here
Video: Find project video here

AIYA Presentation and Workshop

Contributors: Meredith — Archer School for Girls
Overview: A peer-led educational initiative demonstrating that meaningful AI impact comes from teaching critical thinking, not just building software. Meredith designed and hosted interactive student workshops breaking down generative model mechanics, failure modes, and risk evaluation. Participants engaged in hands-on challenges designed to replace passive tool consumption with active, critical audit habits.
Summary: Find project summary here
Video: Find project video here

Carb Counting Tool for T1D

Contributors: Bader — Al-Bayan Bilingual School
Overview: Rooted in personal experience with Type 1 Diabetes, this project leverages LLM capabilities to estimate carbohydrate content from casual meal descriptions—simplifying a daily friction point for T1D patients. Moving from user research into a Dasman Diabetes Institute pilot partnership, the tool’s accuracy is being rigorously benchmarked against standardized nutritional data to ensure safety and precision.
Summary: Find project summary here
Abstract: Find abstract here

Project Erdős

Contributors: Ryan — Saint Stephen's College
Overview: An experimental autonomous multi-agent AI framework designed for continuous, iterative problem-solving on complex, open-ended research questions. Rather than producing one-shot outputs, the system decomposes tasks, executes research, cross-checks candidate solutions across multiple AI architectures, and self-evaluates outputs to systematically improve reasoning quality over repeated cycles.
Summary: Find project summary here
Video: Find project video here