Pitt’s HAIL Awards Put Non-CS PhDs on Responsible AI

Eight University of Pittsburgh doctoral students were honored Sept. 10 for a semester spent writing what responsible AI should mean outside a computer-science lab. The lunch at Posvar Hall closed the first HAIL Graduate Awardee Program, run by the Hub for AI and Data Science Leadership.

Six of the eight work in film, English, history of science, social work, biology, or industrial engineering. Their product was a shared definition, a taxonomy of responsibility principles, white papers, and a journal article still headed for print.

Six of the Eight Awardees Work Outside Computing

Nora Mattern, HAIL’s associate director of responsible AI and data science practices, said the students showed “curiosity and critical thinking about responsibility in AI and data science practices.” She is a teaching associate professor in the Department of Information Culture and Data Stewardship and ran the program as facilitator.

HAIL listed the inaugural HAIL graduate student awardees as the 2025-2026 cohort. Two sit in computing programs. Jennifer Nwogu is in Information Science. Benjamin Panny is in the Intelligent Systems Program. The other six do not.

THE INAUGURAL COHORT

Gatfield-Jeffries, who is beginning her dissertation in the History and Philosophy of Science department, studies AI in scientific research through questions of evidence and ethics. She joined Pitt in 2023 after a science and political-economy degree at the University of Sydney and an MPhil at Cambridge. HAIL’s own spotlight on her work treats those lenses as the point of the award, not a side interest.

HAIL leaders called the Posvar gathering friendly and informal, with a short program on each student and an open stretch to meet faculty and staff from other departments. The mix on the roster is the design. A film student and a social-work student were asked the same two questions as the intelligent-systems student.

They Wrote a One-Sentence Rule for Responsible AI

Mattern charged the group with two questions: what responsible data and AI use and development entail in different disciplines or professions, and what individual actions can put that into practice. Through the spring 2026 semester the cohort met faculty and staff who walked those questions through live research problems, not a generic ethics module.

The students built a taxonomy of responsibility principles and a definition they took to a campus poster session.

Responsible data and AI work integrates human-centered, methodological, and forward-looking actions to balance the risks and benefits of data and AI-driven research.

HAIL Graduate Awardee Program cohort, DataSci+AI Forum, March 2026

That sentence is the portable object. It names three kinds of action and a balancing job, and it does not locate the work in a computer-science department. HAIL can hand it to a lab, a city office, or a hiring manager without translating the jargon first.

THE SPRING SEMINAR

Kitzes also holds HAIL’s associate director post for AI for science, so the reproducibility hour was not a guest lecture from far off campus. Tafti’s explainable-AI session sat next to Mostern’s session on agency. The syllabus treated those as one craft.

HAIL Was Built to Connect Work That Already Existed

The hub did not start from a blank map. Pitt already had data and AI work spread across schools, and HAIL’s own about page says the gap was a way to connect that work and make it visible. Michael Colaresi, a political scientist who had been associate vice provost for data science, was named director of the hub effective Jan. 1, 2026, and strategic advisor to the provost.

Provost Joe McCarthy said the hub would advance the academic mission, support students, network faculty and staff across disciplines, and add to the wider conversation on responsible use and development of AI. HAIL grew out of Responsible Data Science at Pitt, which had already argued that literacy was not enough and that the university needed agency, confidence, and accountability around data work.

THE HAIL CALENDAR

1. December 8, 2025: The provost appoints Colaresi to lead a new academic hub built on the RDS initiative.
2. January 1, 2026: HAIL begins operating as the convening point for AI and data-science work in teaching, learning, and academic operations.
3. Spring 2026: The first graduate awardee cohort meets the seminar leaders and starts the taxonomy.
4. March 26-27, 2026: The students present posters at DataSci+AI Forum, themed The Story and Future of Pittsburgh Through Data.
5. Summer 2026: The cohort writes reflection white papers on a responsibility practice each person can carry into their own research.
6. September 10, 2026: HAIL hosts the celebratory lunch at Posvar Hall.
7. September 17, 2026: Applications close for the 2026-27 Graduate Awardee program.

The graduate program itself is new and sits on top of an undergraduate RDS Student Scholars track. HAIL describes it as an interdisciplinary stint for doctoral and master’s students, with readings and guests on transparency, privacy, and accountability, and a stipend paid on completion. The first eight people through it were doctoral students.

Banks, Hospitals, and Retailers Sit on the Same Board

Colaresi has been explicit that the hub’s outside network is part of the job, not a decoration. In March, ahead of the forum, he pointed to School of Computing and Information colleagues inside Pitt and to the advisory board outside it.

We have such creative and dedicated collaborators internally, such as the School of Computing and Information leadership, and externally, with the 13 companies and organizations that comprise the HAIL advisory board. This dynamic network has built both interpersonal trust and organizational bridges that we believe are necessary to lead in the age of AI.

Michael Colaresi, HAIL director, March 2026

HAIL’s published board list puts regional finance, health systems, retail, logistics, chemicals, government, and philanthropy in the same room. The named organizations include First National Bank, PNC, Highmark, UPMC, Dick’s Sporting Goods, Accenture, the City of Pittsburgh, the Hillman Foundation, FedEx, PPG, BNY, and Bluestreet Data, whose Andrew Hannah chairs the group.

WHO HAS A SEAT

• Health systems: UPMC and Highmark send data and analytics leaders, including UPMC vice president Suresh Mulukutla and Highmark chief analytics officer Richard Clarke.
• Banks: PNC, BNY, and First National Bank sit alongside the health names, the same employers that hire Pitt graduates into analytics roles.
• Retail and logistics: Dick’s Sporting Goods, FedEx, and PPG add store, package, and manufacturing data to the brief.
• Civic and nonprofit: The City of Pittsburgh’s Chris Belasco and the Hillman Foundation’s Meredith Mavero keep a public-sector line into the hub.

Kendra Oliver, HAIL’s managing director, told a Dietrich School graduate series in February that the board’s value was two years of talk about what those employers want in new hires. Technical skill still matters, she said, and then she named the other list: critical thinking, collaboration, and creativity. The graduate awards put people from English, social work, and history of science into that hiring conversation on purpose.

HAIL now runs four boards, for industry, academia, research, and students, and it is standing up a Student Advisory Board Committee this year. The graduate awardees are one feeder into that student voice. They are also a feeder into the external board’s workforce brief.

What the March Forum Was Selling

The posters were not a closed-door class show. DataSci+AI Forum ran March 26 at the William Pitt Union and March 27 at the Heinz History Center, under the title The Story and Future of Pittsburgh Through Data. HAIL says the two days brought together over 350 students, faculty, industry, government, and community leaders.

The pitch on the page was that Pittsburgh and the university that shares its name can host the latest digital-technology turn, after the city’s earlier shift from steel into universities and hospitals. Day one was aimed at Pitt leadership, faculty, students, centers, and labs. Day two brought in external partners and community groups. Colaresi gave opening remarks. Hooman Rashidi, associate dean of AI in medicine, keynoted.

Pitt’s School of Computing and Information posted from the first day as faculty, staff, and students moved through the union.

We had an amazing time at day one of Pitt's 2026 DataSCI+AI Forum! 💙 SCI faculty, staff, and students connected with the broader Pitt community to explore how they can innovate and empower through data science and AI tech. See you tomorrow for day two!💡https://t.co/HgjyXCbdgX pic.twitter.com/jD0nUmuK2U

— School of Computing and Information (@pitt_sci) March 26, 2026

The forum is also where the responsible-AI sentence left the seminar and met regional employers. Students got a look at applied AI. Faculty were asked to tie their work to civic problems. Partners were asked to talk workforce. The awardee taxonomy sat in that mix as a student-made object, not a policy memo from the provost’s office.

The same campus is already living with generative AI in a much messier way. A Pitt Digital-sponsored project, GenAI Conversations, sat down with 95 undergraduates across four campuses in spring 2025. Lead researcher Annette Vee, an associate professor of English, and colleagues reported an estimate that 90% of students use generative AI tools, including in courses that ban them, and that most of those students want clearer guidance rather than a ban or a free pass. HAIL’s doctoral cohort is writing practice for labs and community projects while undergraduates are still guessing the rules in class.

White Papers That Changed How the Students Work

After the posters, HAIL asked each awardee for a white paper that reflected on the program and named a responsibility practice they would carry into their own data and AI work. Mattern said those papers are the handoff from the seminar to the dissertation, the field site, or the shop floor.

WHAT THE PAPERS COMMITTED TO

• Documentation: Two awardees wrote that talks on reproducibility and replication pushed them to build extra documentation into their research, the unglamorous half of making a result checkable.
• Shared control: A student whose work designs technology with local community groups wrote that sessions on data stewardship, respect for people, and accountability led her to give collaborators more say in how data is shared and managed.
• A journal record: The full group also drafted a short article on the program’s structure and outcomes, set for a fall 2026 issue of the Journal of Interactive Technology and Pedagogy.

Those three outputs travel farther than a lunch program. A documentation habit follows a student into the next experiment. A community-control habit follows a designer into the next partnership. A pedagogy journal article follows the program into other universities looking for a syllabus they can copy. HAIL already measures the awardee track among the ways it counts “direction setting” work, next to an AI Playbook and teaching field notes.

None of that settles how a bank, a hospital, or a logistics firm will actually ship a model. It does give those firms a Pitt-stamped vocabulary and a small class of doctoral students who have rehearsed it in public.

Applications for the Next Cohort Close Sept. 17

HAIL is already recruiting for 2026-27. Pitt graduate students can apply through the hub’s online form through Thursday, Sept. 17, seven days after the Posvar lunch. The posted program is open to doctoral and master’s students, with the same core questions about what responsible practice looks like in a given field and what one person can do about it.

The first cohort’s definition, taxonomy, and white papers are now the baseline the next group will walk into. The advisory board is still the audience on the other side of campus. And the people being asked to write the practice, if this year is the template, will keep coming from English and social work as much as from intelligent systems.

Frequently Asked Questions

What Is the Hub for AI and Data Science Leadership?

HAIL is Pitt’s academic convening point for AI and data-science work, housed with a strategic advisor to the provost. It uses a method it calls Responsible Data and AI Leadership, or ReDAIL, and a searchable campus map called the Rol-AI-Dex, and it answers through four boards covering industry, classrooms, research, and students rather than through a single ethics office.

What Undergraduate Track Did the Graduate Awards Grow From?

The RDS Student Scholars Program is HAIL’s undergraduate counterpart, run with the Sara Fine Institute in the School of Computing and Information and the Research, Ethics, and Society Initiative in the Office of Research. Scholars meet across a year, build responsibility-focused projects, present at Data Science and AI Day, and also receive stipends.

What Did Rose Gatfield-Jeffries Study Before the HAIL Award?

Before the Pitt doctorate she completed a Bachelor of Science and Advanced Studies at the University of Sydney in History and Philosophy of Science and Political Economy, then an MPhil in History and Philosophy of Science and Medicine at the University of Cambridge. Her current interests include evidence robustness, the role of values in scientific practice, and epistemic injustice in medicine.

Who Leads HAIL Day to Day Besides Michael Colaresi?

Kendra H. Oliver is managing director. Justin Kitzes, a biologist, is associate director for AI for science. Nora Mattern leads responsible practice. Oliver’s own training is a pharmacology PhD plus a UX design degree, a pairing she has said matches the hub’s brief to make AI usable across campus rather than only inside computing.