Publications

A growing record of HVP research advancing the alignment of AI with human values in medicine

The Missing Dimension in Clinical AI: Making Hidden Values Visible
NEJM AI

The Missing Dimension in Clinical AI: Making Hidden Values Visible

Goldberg C, Balicer RD, Bhat M, Blumenthal D, Brendel RW, Brondolo E, Brownstein JS, Buckley T, Cain CH, Chandak P, Chessa F, Chopra A, Dagan N, Ehlert KS, Evans BJ, Freeman R, Glicksberg BS, Gordon W, Gröschel MI, Hoffman S, Hundert E, Hyare S, Johri S, Joseph J, Klote M, Landman AB, Lee VS, Mandel JC, Mandl KD, Manrai AK, Might M, Nadkarni GN, Nigrin DJ, Noori A, Omenn GS, Parimbelli E, Rosenberg AL, Stutz D, Tory M, Tunik E, Wolf SM, Zitnik M, Kohane IS

NEJM AI, 2026

Consensus statement from the RAISE symposium proposing a 'Values In the Model' (VIM) framework — a transparent labeling system that documents how AI systems navigate value-laden clinical trade-offs, enabling informed choices about AI alignment.

View PDF DOI View
A Global Log for Medical AI
arXiv

A Global Log for Medical AI

Noori A, Rodman A, Karthikesalingam A, Mateen BA, Longhurst CA, Yang D, deBronkart D, Galea G, Wolf HF III, Waxman J, Mandel JC, Rotich J, Mandl KD, Mustafa M, Miles M, Shah NH, Lee P, Korom R, Mahoney S, Hain S, Wong TY, Mundel T, Natarajan V, Dagan N, Clifton DA, Balicer RD, Kohane IS, Zitnik M

arXiv, 2025

Proposes a global, standardized logging framework for medical AI systems to enable transparent tracking of AI deployment, performance, and impact across healthcare institutions worldwide.

View PDF DOI View
The Missing Value of Medical Artificial Intelligence
Nature Medicine

The Missing Value of Medical Artificial Intelligence

Kohane IS, et al.

Nature Medicine, 2025

Argues that the critical missing element in medical AI development is systematic measurement and incorporation of human values into system design and evaluation.

View
The Human Values Project
ML4H Symposium

The Human Values Project

Kohane IS, Chandak P, et al.

ML4H Symposium, 2025

Introducing the Human Values Project framework and initial findings from studying how 15 AI models handle ethical dilemmas in clinical triage scenarios.

View PDF View
Compared with What? Measuring AI against the Health Care We Have
NEJM

Compared with What? Measuring AI against the Health Care We Have

Kohane IS

NEJM, 2024

An argument for evaluating medical AI not against theoretical perfection, but against the actual quality and consistency of current clinical practice.

DOI View
Medical Artificial Intelligence and Human Values
NEJM

Medical Artificial Intelligence and Human Values

Yu KH, Kohane IS

NEJM, 2024

A comprehensive review examining how AI systems encode and express values in medical decision-making, and why aligning these systems with human values is essential for healthcare.

View PDF DOI View
To Do No Harm — and the Most Good — with AI in Health Care
NEJM AI

To Do No Harm — and the Most Good — with AI in Health Care

Goldberg CB, Adams L, Blumenthal D, Brennan PF, Brown N, Butte AJ, Cheatham M, deBronkart D, Dixon J, Drazen J, Evans BJ, Hoffman SM, Holmes C, Lee P, Manrai AK, Omenn GS, Perlin JB, Ramoni R, Sapiro G, Sarkar R, Sood H, Vayena E, Kohane IS, for the RAISE Consortium

NEJM AI, 2024

Consensus from the inaugural RAISE conference calling for AI in healthcare to enhance patient care, support clinicians, and be accessible and safe for all, with specific actions for health care leaders.

View PDF DOI View