ArtInsight
AI-powered system to facilitate deeper engagement with child-created artwork in mixed visual-ability families.

Image showing a number of components I created during the internship
Image showing a number of components I created during the internship

Timeline

Summer 2024 – Spring 2025

Roles

Interaction Designer

User Researcher

Team

Arnavi Chheda-Kothary

Ritesh Kanachi

Kevin Xiao

Aditya Sengupta

Chris Sanders

Tools

Figma

Figjam

Premiere Pro

Overview

Context

ArtInsight is a project under the Makeability Lab, an applied machine learning lab focused on accessible technology.

My Role

As a designer and researcher, I created interfaces reflecting key insights from an initial study using the IOS 18 UI kit, assisted with usability testing and thematic analysis, and created video demonstrations for conference submissions.

Impact

Paper accepted to ACM IUI 2025.

Challenge

We built on prior research on the positive impact of AI art description in mixed visual ability families. Key findings included the desire to tweak AI outputs and that descriptions should be respectful and thorough.

Current image-to-text applications intended to assist blind or low-vision individuals lack artistic language and nuance.

Opportunity

We developed ArtInsight using Figma and SwiftUI, powered by a custom Chat-GPT 4o assistant.

Features

AI Descriptions

Users can toggle between a more objective, descriptive description and a more presumptive, creative description for varying perspectives.

Audio Recording

Allows children to record a voice memo to describe the artwork themselves, which may be used to feed additional context into the artwork description.

Questions

Generates set of three example questions, used to stimulate and strengthen discussion.

“You want your child to feel like you’re acknowledging their work… [With ArtInsight], I feel more truly connected to the drawing.” – P4 

“You want your child to feel like you’re acknowledging their work… [With ArtInsight], I feel more truly connected to the drawing.” – P4 

Case Study In Progress

© Chris Sanders 2025
© Chris Sanders 2025
© Chris Sanders 2025
© Chris Sanders 2025
© Chris Sanders 2025
© Chris Sanders 2025