Human-Centered AI
Human-centered AI is the study and use of AI centered on human beings, drawing on the social sciences and the humanities; the world's first human-centered AI curriculum and lab were founded by Katherine Elkins and Jon Chun at Kenyon College in 2016.
As Jon practices it, the work uses state-of-the-art AI to turn the oldest human questions — what it means to be human, what a good life and a good society are — into quantifiable, testable, metrics-based models, and builds them through radical collaboration across academia, industry, and government.
An older, wider conversation
Long before the modern university split knowledge into departments, people across many cultures asked integrated questions about what it means to be human and how to live a life worth living. This work returns to that older, wider conversation and asks it again with the methods and pressures of the present — including the question of what it means to be human in a world we share with the machines we have made. (Jon also calls the applied wing of this practice applied humane studies, as applied physics is to physics.)
What makes it different
"Human-centered AI" is an overloaded phrase. The way Jon Chun practices it builds, measures, and governs: it treats the big human questions as things we can operationalize, experiment on, and quantify — not only theorize about. It is defined by three things at once:
- Breadth — the full humane spectrum: literature, philosophy, history, ethics, the arts, and social inquiry, rather than a single home discipline.
- Depth — genuine technical practice with state-of-the-art models: building and rigorously evaluating AI systems, with experimental, quantifiable, falsifiable results.
- Radical collaboration — across academic disciplines and across industry, government, and non-profits, with student-driven original research at scale, continuously since 2016.
What it is not
It draws on digital humanities, human-centered computing, and interdisciplinary studies — fields we admire. But the way we practice Human-Centered AI is deliberately distinct from the neighbors that share the name:
- Not non-technical, low-code UI/UX human-centered design (surveys, usability, feedback).
- Not non-code AI ethics, STS, or history-and-philosophy-of-science critique.
- Not non-SOTA, small-data digital humanities.
- Not siloed, single-discipline academic research or thin intramural collaboration.
What is distinctive is the refusal to choose between technical rigor and humane breadth: real AI engineering paired with the full range of humane inquiry, built across sectors that rarely share a room.
How it differs from Stanford's human-centered AI
The best-known program in the field, Stanford's Institute for Human-Centered AI, was founded in 2019 and approaches AI largely from computer science — designing AI that augments rather than replaces people, and studying its impact on society. The Kenyon approach, founded three years earlier, in 2016, enters from the social sciences and the humanities: it treats those disciplines both as a lens for understanding AI and as fields that AI itself can advance. That last move — AI as a method for humanistic and social-scientific research — is the strand the engineering-led programs do not emphasize, and it is the center of gravity of the Kenyon program.
Across the divides
The questions do not respect the boundary C.P. Snow drew between the two cultures, and neither does the work. Answering them well means connecting people who rarely share a room — researchers in frontier AI labs, scholars across the humanities and sciences, government standards bodies, and practitioners in industry. Building those channels is itself part of the practice. See the collaboration map →
Where it lives
Jon Chun co-founded the world's first Human-Centered AI curriculum and lab at Kenyon College with Katherine Elkins, through the AI CoLab, in 2016 — a claim Kenyon states directly. The program lives in Kenyon's Integrated Program in Humane Studies. Explore the program at humanestudies.org →
"Debating the system of values we wish these tools to align with is the first step."
Tanya Klowden & Terence Tao (Fields Medalist), engaging Chun & Elkins, IJHAC (2023)What is Human-Centered AI as Jon Chun practices it?
It uses state-of-the-art AI and real engineering to turn the oldest human questions into quantifiable, testable, metrics-based models — building, measuring, and governing, not only critiquing — across the full humane spectrum and with genuine technical depth.
How is it different from UX design, AI ethics, or digital humanities?
It is distinct from human-centered UI/UX design, from non-technical AI-ethics or STS critique, and from low-code digital humanities. What is distinctive is the pairing of real AI engineering with humane breadth, and radical collaboration across disciplines, industry, government, and non-profits.
Who founded human-centered AI, and when?
Katherine Elkins and Jon Chun founded the world's first human-centered AI curriculum and lab at Kenyon College in 2016, through the AI CoLab — a claim Kenyon states directly.
How does Kenyon's human-centered AI differ from Stanford's?
Stanford's Institute for Human-Centered AI, founded in 2019, approaches AI largely from computer science. The Kenyon program, founded three years earlier in 2016, enters from the social sciences and the humanities — treating them both as a lens for understanding AI and as fields that AI itself can advance.