Co-Creative AI
For this project, I designed and implemented the interaction design and system design on my own using a mixed-method approach, including developing personas, user journeys, paper prototypes, and mockups. The Drawing Partner is a co-creative drawing agent that analyzes the user’s contribution, feedback, and instructions to produce a creative contribution on a shared canvas. The system is loosely turn-based, but the user can draw at the same time as the agent, which makes synchronous collaboration possible. The system is also a quantified co-creative AI system that quantifies and models co-creation based on the creative sense-making framework. The AI Drawing Partner is a publicly available web app (https://drawing-partner.vercel.app).
In this project, I was the user experience lead and project lead. The Drawing Apprentice is an early co-creative drawing partner that can collaborate with users in real time on a shared canvas. It analyzes the user's lines and generates its own responses based on the user's input and previous feedback. It uses a combination of machine learning algorithms to recognize the user's sketched object, as well as their positive and negative feedback through time. The system is meant to engage users in a creative dialogue to help inspire new ideas, creatively engage the user, and emphasize the creative process over product (lowering the barrier of entry for novices). The Drawing Apprentice serves as an experimental platform to explore the technical approaches and interaction designs that help facilitate co-creation. It helps answers questions about what it means to collaborate with a creative computer and how people ideally imagine co-creation with a computer.
Helped project leads Mikhail Jacob and Duri Long analyze and evaluate the user experience and interaction design of the co-creative dance agent and LuminAI dome exhibit. This project enables users to dance with virtual agents that respond to the user's dance moves in real time. This project explores interactive machine learning and performance based co-creative systems.
© Nicholas Davis 2022