About

Hyeongdon (Doni) Moon is a PhD student in the Human-Computer Interaction Institute at Carnegie Mellon University, advised by John Stamper and Carolyn Rosé. He holds an M.S. in Computer Science from EPFL and a B.S. in Computer Science and Engineering from Korea University.

Before his PhD, he worked at Riiid in Seoul (2020–2023), first as an AIOps back-end engineer, then as a research scientist building an AI-powered TOEIC tutor and doing educational NLP and multimodal-content research.

His research interests are AI-enhanced knowledge measurement, motivated by ethical concerns about how society evaluates people, and artificial consciousness, drawing on his background in cognitive science, NLP, and neuroscience.

His current PhD work centers on modeling how humans acquire knowledge — building cognitive-architecture-inspired representations of problem solving, and extending knowledge tracing so learning itself can be measured and supported more directly.

Doni's essay, Embrace and Striving: At the Frontier of Technology in the AI Era, starts from the fact that human understanding is finite, and argues that the rational response to that finitude is acceptance — loving your own limits unconditionally — paired with striving, the ongoing effort to expand your world and meet real threats. It treats these as complements, not opposites: accepting your limits is what makes sustained striving possible instead of self-destructive.

It extends that view outward as a practical humanism — reasoning from behind a veil of ignorance about which mind you might have ended up being — and applies it to the present moment, arguing that AI is eroding the one lever ordinary people held over politics and the economy, and that this is a crisis for democracy in need of deliberate redesign.

Doni's proposed response is what he calls a flexible information society: decision-making systems that keep pulling ordinary people's knowledge and judgment back into the loop, rather than letting efficient top-down governance route around them.

Opinion Interface is the platform he built to test this — each person's digital AI persona answers questions on their behalf, grounded in their own public statements and writing, and forwards on whatever it can't yet answer.