The PHLAIR Fellowship — Public Health Leveraging AI Responsibly | Boston University School of Public Health

About the Fellowship

Mentored research at the intersection of AI and public health

AI offers public health some of its most promising opportunities in a generation: faster research, earlier detection of threats, smarter use of scarce resources, and new ways to reach people whom services have historically missed. Realizing that promise responsibly requires practitioners who pair technical fluency with public health values — who pursue AI's benefits energetically while weighing its real costs in ethics, equity, and ecological impact.

The PHLAIR Fellowship — the BUSPH Fellowship for Public Health Leveraging AI Responsibly — is sponsored by Boston University School of Public Health (BUSPH) and gives MPH students a structured, mentored, and paid 100-hour research experience — starting the term after completing the Core (typically spring) — culminating in a research poster. Fellowship participants are selected from accepted students in BUSPH's in-person MPH program.

Fellows work with faculty mentors drawn from across BUSPH's six academic departments and the Center for Health Data Science, alongside seminars on responsible AI and leadership in public health.

At a Glance

Who: Fellowship participants are selected from accepted students in the BUSPH in-person Master of Public Health (MPH) program
Not yet at BUSPH? Apply to the in-person MPH program — acceptance is the first step toward the fellowship
What: A mentored research project in one of the fellowship's AI-and-public-health focus areas
When: 100 hours of paid work, starting the term after completing the Core (typically spring)
Outcome: A research poster presented at the MS/MPH Student Poster Presentation Day

The Program

Fellowship components

Fellowship work is supported by a stipend set on an annual basis, expected to be $2,000 for the incoming cohort of fellows. All fellows are expected to attend an introductory orientation meeting outlining the major challenges and opportunities in AI and public health and discussing core skills. At this time, fellows will learn about and rank the specific projects in each faculty mentor's team. Fellows will be matched to a mentor based on ranking and faculty availability. Mentors typically have 1–2 fellows, who join their larger research team — which may include PhD students, postdoctoral fellows, and other scientific and programmatic staff.

After the introductory meeting, fellows select at least three more seminars to attend during their fellowship, aligned with their specific interests within the broad domain of AI and public health. These may include Epi-Informatics hands-on workshop activities, visiting speaker seminars, or specific skills trainings available to BU students.

BUSPH students working together around a conference table with laptops during a weekend session — BUSPH students during a session extracting data to train an AI algorithm used to evaluate p-hacking. See Professor Glymour's post on LinkedIn →

PHLAIR fellows don't work alone. Joining a mentor's research team means working alongside PhD students, postdoctoral fellows, and staff — in sessions like this one, where ideas, code, and questions move around the table. As fellowship contact Maria Glymour shared on LinkedIn: there's nothing better than a Saturday afternoon spent doing research together.

PART 1

AI-Focused Seminars and Trainings

An introductory orientation plus at least three seminars chosen by each fellow — Epi-Informatics hands-on workshops, visiting speaker seminars, or specific skills trainings available to BU students.

PART 2

Mentored Research Project

The core of the fellowship: a paid, independent project in one of the focus areas, scoped to roughly 100 hours, starting the term after completing the Core, and guided by a BUSPH faculty mentor as part of the mentor's research team.

PART 3

Poster & Presentation

Fellows prepare a research poster and present it at the MS/MPH Student Poster Presentation Day, joining student fellows from programs across BUSPH.

The Fellowship

Focus Areas

AI's promise for public health is real — and so are its costs. Each fellow anchors their project in one or more of the fellowship's focus areas, identifying the most promising strategies to empower public health practitioners and leaders.

No. 01

Ethics of AI in Public Health

Consent, transparency, accountability, and governance when algorithms inform population-level decisions — from surveillance and screening to resource allocation.

No. 02

Ecological Impact of AI

AI is both a tool for research on ecological and environmental health and a contributor to emissions through energy- and water-intensive computing. Fellows examine the ecological footprint of health AI and AI's applications for ecological adaptation and resilience.

No. 03

AI to Accelerate Research

From literature synthesis and study design to natural language processing of health records and disease forecasting — methods that shorten the path from question to evidence without sacrificing rigor. Fellows also examine how AI affects the reproducibility of science and its growing role in meta-research: using AI to study, synthesize, and improve research itself.

No. 04

Fairness & Bias in AI

Models trained on biased data can encode and amplify health inequities. Fellows audit algorithms for differential performance and study how biased AI affects marginalized communities.

No. 05

Applications for Practice — Improving Access to Services

The ultimate test of AI in public health is whether it improves lives. Working in sustained partnership with community organizations, health departments, community health centers, and health systems, fellows explore real-world applications: multilingual tools that connect residents to benefits and care, decision support for under-resourced agencies, and outreach systems that reach people whom services have historically missed. Community partners help shape these tools from priority-setting through design and evaluation.

No. 06

AI to Improve Health Communications

Clear, trusted communication is core to public health practice. Fellows explore how AI can help tailor health messages to different audiences, languages, and literacy levels; support practitioners in detecting and responding to misinformation; and help health departments communicate quickly, accurately, and accessibly — including during emergencies.

Share Your Research

Present at MS/MPH Student Poster Presentation Day

The fellowship culminates in a research poster presented at BUSPH's annual MS/MPH Student Poster Presentation Day, where student fellows from programs across the School share original research conducted under the mentorship of faculty and community partners.

To see examples of fellows' presentations, read the SPH News story on this year's event, featuring student fellows from fellowships across the Department of Community Health Sciences, the Department of Epidemiology, and the Center for Trauma and Mental Health:

Read: "Student Fellows Present their Original Research" (SPH News) →

Student fellows present research posters at the BUSPH MS/MPH Student Poster Presentation Day — Student fellows present their research at MS/MPH Student Poster Presentation Day. Photo: Megan Jones, BUSPH — from the SPH News article linked at left. Confirm image use with SPH Communications before launch.

Example Projects

What a PHLAIR project can look like

Two illustrative examples of the kinds of projects fellows could pursue — one community-engaged, one focused on AI's role in research itself.

Community engagement for AI in health and healthcare

Responsible AI in health is built with communities, not merely for them. In this example, a fellow partners with a Boston-area community health center to assess how an AI-assisted, multilingual "health navigator" tool could help residents find and enroll in services they already qualify for — primary care, behavioral health, food assistance, housing support, and transportation to appointments.

Working with the health center and the residents it serves, the fellow might conduct listening sessions to identify the real barriers to accessing services, draft a community data-governance plan, or audit an existing tool's performance across languages and neighborhoods — reporting results back to the community in plain language.

Evaluating AI-generated summaries of scientific articles

Academic journals report that readers are increasingly relying only on AI-generated summaries instead of clicking through to read the original article (or even the abstracts). How well does AI summarize the health and medical literature — and can readers tell the difference? In this example, a fellow selects a handful of manuscripts published in health and medicine, writes human-generated summaries, and elicits AI-generated summaries of the same articles.

Independent, masked readers — blinded to whether each summary was written by a human or an AI — then evaluate the summaries for clarity, accuracy, and overall quality. The results speak directly to AI's role in accelerating research, science communication, and the reproducibility of evidence synthesis.

Faculty Research Spotlight

Mary D. Willis, PhD, MPH

Dr. Mary Willis is an Assistant Professor in the Department of Epidemiology at BUSPH. Her work sits at the intersection of environmental epidemiology, spatial exposure assessment, and applied data science, and frequently uses econometric-based causal inference methods — a model for the kind of rigorous, data-intensive, policy-relevant research this fellowship trains students to do.

Her research examines how exposures from the energy sector — such as oil and gas development and traffic-related air pollution — and features of the built environment, including green space and neighborhood disadvantage, influence reproductive health. She is principal investigator of an NIH Director's Early Independence Award examining the health impacts of oil and gas development, and she led a BUSPH study, published in the American Journal of Public Health using the BUSPH-based Pregnancy Study Online (PRESTO) cohort, that found people trying to conceive who lived near active oil and gas development had a heightened risk of moderate-to-severe depressive symptoms. She also helped launch a database with an SPH team showing which communities are most affected by fossil fuel sites.

View Dr. Willis's BUSPH profile →

Why her work matters for fellows

Applied data science and spatial methods on large, real-world datasets
Causal inference designed to inform health-protective policy
Direct relevance to the fellowship's ecological impact focus area
Research that translates data into protections for affected communities

A fellowship that shapes how you'll use AI for the public's health.

PHLAIR fellows are selected from accepted students in the BUSPH in-person MPH program.

Not yet a BUSPH student? Apply to the in-person MPH program first →

MPH application deadlines

September 2026 entry: the final deadline for domestic applicants to the on-campus MPH is June 15, 2026 (the international deadline was April 1, 2026). Applications must be submitted by that date.

January 2027 entry: BUSPH reviews applications on a rolling basis for spring matriculation through SOPHAS. The Spring 2027 deadline will be posted on the BUSPH applying page; in past cycles the domestic spring deadline has fallen in mid-December. Questions: asksph@bu.edu.

Questions about the PHLAIR fellowship? Contact Maria Glymour, Professor and Chair, Department of Epidemiology, BUSPH

The PHLAIR Fellowship.