What I work on

We are living through a moment when the line between human agency and machine agency has never been more blurred. As AI moves into decisions once made by people alone, a basic question keeps getting lost: what is still ours to do, and what have we quietly handed over?

I believe in human-centered AI— that AI was built, first and foremost, to serve people, to free us from meaningless work, not to replace the parts of work that make us who we are. Humans are an extraordinarily adaptive species; we will always find a way to work alongside whatever tools we’re given. The real question is what kind of collaborator we are building, and what that collaboration does to us.

My research tries to answer this by watching where people look. Eye movement is a natural carrier of attention — and attention doesn’t just reflect cognition, it shapes behavior. By combining eye-tracking with process-oriented behavioral data, I study how people actually interact with AI in real time: when they trust a recommendation, when they override it, and why. This is not only a way to understand humans. It is a way to understand what the next generation of AI needs to become in order to collaborate with us well.

I ground this work in high-stakes domains — electricity demand forecasting and air traffic control — where the cost of getting human-AI collaboration wrong is measured in real consequences, and where neither blind trust nor blind distrust is safe.

What I hope to contribute to is human-AI interaction and alignment — not only as a safety problem, but as a question of symbiosis: how humans and AI can genuinely work together, each doing what they do best, without either side losing what makes it valuable.