Hackathon Winner: Helping teachers adapt their lesson plans to physical and emotional state of students using generative AI

Overview

Popsicle was born from a desire to help teachers understand their students better, enabling them to tailor their lessons to individual student needs with the help of AI. During the Code Your Dream's accessibility-themed Hackathon, I collaborated with a team of passionate educators to address this accessibility challenge. Our goal was to empower teachers to plan lessons that consider the unique needs of all students, regardless of neurodiversity or other factors.

Challenges and goals

Teachers lack the knowledge to address individual student needs, including neurodiversity, limiting an optimal learning environment. Privacy concerns hinder personalization. With limited time, we swiftly assigned roles, defined the problem, gathered feedback, researched AI tech, and designed a solution. Our goal was to provide personalized knowledge to teachers daily and offer students suggestions on dealing with personal problems for better learning. We want to empower teachers and students to create a supportive and inclusive environment, helping every student reach their full potential.

Process and role

As Project Manager and Product Designer, I led the process, from problem definition to prototyping, ensuring a comprehensive solution for teachers.

Research

To address the challenges faced by teachers in understanding and supporting both neurodiverse and regular students, we interviewed our teacher teammates to uncover their pain points. While conversations were helpful for regular students, additional factors like food availability, household dynamics, general readiness, and neurodiversity posed complexity. Some students were unable or unwilling to communicate their problems. We sought a solution that respects student privacy while gaining insights into their well-being, allowing teachers to create a supportive learning environment that enhances student learning.

Solution

Large Language Model (LLM)- an artificial intelligence (AI) system that understands and generates human language. It's trained on massive amounts of text data to perform tasks like answering questions and generating text. It's comparable to a super-smart language assistant.

Choosing simplicity amidst emerging tech trends, I opted for an LLM to analyze student survey responses—covering daily feelings, readiness, food, and home issues. The LLM then offers insightful tips based on its trained knowledge.

Our solution includes a teacher dashboard, individual student notifications, and a log. To address privacy concerns, we offer daily data, a privacy agreement page, and let students choose if teachers can view responses. This shapes our Popsicle project, empowering teachers while respecting student privacy.

Design

Considering the feedback from our teacher teammates, most classes relied on a single computer for class software. With a tight timeframe, we opted for a responsive design, prioritizing a mobile-first approach. By starting with the smallest screen, we focused on essential elements and scalability. To create a visually engaging experience, we chose an orange primary color to evoke security, warmth, and happiness, while utilizing vibrant colors in the survey to keep students engaged and convey emotions.

Outcome

In just two hours, our team researched, ideated, and designed the product. We presented a prototype to a panel of four judges and won the Hackathon, receiving the "Most Innovative" award for using LLMs to address academic challenges.

Takeaways

Given more time, we would have conducted broader user interviews to gain insights from teachers and students. Usability testing would have been valuable to gather feedback on the prototype. Additionally, measuring student grades and teacher perception of progress would provide real-world impact assessment.

View more work

Check out other case studies or articles