I’m curious about how we can use AI to connect more deeply with students’ unique experiences, helping them explore their interests and make progress toward their goals. It’s a big, messy question, but one I find really exciting to think about.

Right now, I’m exploring three related questions:

What can current AI actually do well, and where does it fall short? I’m particularly interested in whether large language models (LLMs) can reflect the diversity in how students explain scientific ideas and how well they can hold onto and recall details about a student’s experiences to offer personalized support.

How can we design AI to make the most of its strengths? I’ve been thinking a lot about anthropomorphism—whether we should design AI to feel like tools or more like human-like agents—and what difference that makes in how people use them.

How do these ideas hold up in real classrooms? I’m trying to understand how students perceive AI’s personality and intentions and how that shapes their interactions with it. Does it feel supportive? Frustrating? Something else entirely?