"I envision a mixed-initiative co-creative future where AI is no longer the imposer of norms and biases on human-AI collaborative relationships ..."
I started diving into computer science when I was a primary school student. This was rare to my community back in China where we count "schools per computer". I have since deeply cultivated my interest in programming and software engineering, which eventually lead me towards pursuing a Software Engineering degree at Shanghai Jiao Tong University via a special recommendation for winning the first prize in a provincial programming competition.
At the same time, I started playing a lot of video games and volunteered in developing them. Even with my programming expertise, I realized I still lacked the knowledge needed to create all components of a video game by hand. Furthermore, building content for games was (and still is) challenging and time-consuming, even for the domain experts.
With these insights, I joined Georgia Institute of Technology where I met Professor Mark Riedl. Together, we explored AI, specifically neural networks, as part of revitalized technology. In 2016, we studied Procedural Content Generation trade of computer generated game content. We quickly realized that computers themselves were not enough for generating this kind of content. Without creator-awareness, whatever being generated remains unrelated to creator needs. Along with other research within the realm of Reinforcement Learning and Human-Centered AI, I spent the rest of my PhD journey diving into mixed-initiative co-creativity where human creators and AI share contexts and initiatives collaborating in creative activities.
As I finish my PhD journey, I began to realize how challenging it is for both humans and AI to understand each other. Even with co-creativity's potential, if either party cannot effectively understand the other, nothing creative will yield. This is where UCSC's focus on games and computational creativity and GUII Lab's unique combination of projects in games, education and affective science drew me in and allowed me to deepen my understanding of human-AI collaboration.
I study how humans and Artificial Intelligence work together, learn from each other, and go beyond each other alone.
Research in AI Procedural Content Generation and my deep interest in game and software engineering led me towards the gap between opaque AI and human needs: Human-Centered AI.
Perhaps my research in rhythm action game stage generation "GenerationMania." I was a high level player and created levels that millions played. I believe I was also the first one intersted in generating those specific types of rhythm action game stages. This was also the topic of my first accepted paper.
I envision a mixed-initiative co-creative future where AI is no longer the imposer of norms and biases on human-AI collaborative relationships and that we are able to achieve true human-AI collaboration.
Specifically within the domain of games, I envision that this will truly level the playing field for game development, to the point where "game development experts" no longer truly exist. Anyone interested in game development would be able to collaborate with an AI agent to envision and build whatever games they wish to play. They would no longer be limited to labor and cost considerations.
There are no "stupid questions." If you wish to learn something new, something that no one has researched before, the door is always open to you!
This is where it all started :)