GW Takes Major Step toward Comprehensive AI Strategy with University-Wide Mapping Exercise

Authored by:

George Washington University President Ellen M. Granberg announcedWednesday that GW is developing a comprehensive, aligned strategy for AI as an essential supporting pillar of the university's strategic framework. Leading that effort as senior advisor to the president on AI strategy is Zoe Szajnfarber, a professor of engineering management and systems engineering with secondary and courtesy appointments in international affairs and computer science. Szajnfarber also is faculty director of the GW Trustworthy AI Initiative(GW TAI).

"I want to thank Dr. Szajnfarber for bringing her expertise to this critical leadership role that will help guide the university toward a comprehensive, dynamic and mission-aligned AI strategy," Granberg said. "Leveraging the critical work that is already underway across the entire university and anticipating future directions in this rapidly-changing space, this project touches all three priorities of our Strategic Framework and-like the framework-will require the entire community to engage and contribute so that our AI strategy reflects who we are as an institution and where we are heading."

Over the coming months, Szajnfarber will spearhead a university-wide strategic mapping exercise on the uses, opportunities and risks of AI in every aspect of GW's institutional processes. The exercise will explore how these technologies are already being used-and where new challenges and paths forward have emerged as a result-across three key pillars: research, teaching and operations.

The team has already solicited feedback from elected faculty representatives, including the co-chairs of the Faculty Senate's committees on research and on educational policy and technology, and is now in the process of assembling working groups with representation from across the GW community. The research working group will be co-chaired by David Broniatowski, professor of engineering management and systems engineering at GW Engineering, and Ted McKoy, assistant director for award set-up and subaward management in the Office of Sponsored Projects; the education working group by Ryan Watkins, professor of educational technology leadership in the Graduate School of Education and Human Development, and Karen Singer-Freeman, associate dean of teaching and learning excellence in the Center for Teaching Excellence; and the operations working group by Sue Bogan, GW IT digital transformation specialist, and Patrick Hall, teaching assistant professor of decision sciences and chief AI officer in the School of Business.

These groups will collect and analyze data from schools, units and the community at large, and their results will provide the foundation for GW's institutional AI strategy, which will serve as a resource for academic departments, schools and administrative divisions as they consider how and where AI should be implemented across curriculum, research and university operations. (Specific details on how to join these conversations are forthcoming. Community members interested in participating can register interest by filling out this form.)

"It's hard to imagine any discipline, topic area or organizational unit not touched by AI in some way," Szajnfarber said. "As we've had the chance to talk to colleagues across campus about their AI experiences, I've been struck by the range of apprehension and excitement, but also the nuances in all the different ways people are leveraging it in their contexts. I've already refined my thinking so much through those interactions."

Szajnfarber's interdisciplinary background and years of experience with AI, both generally and at GW specifically, make her the ideal leader for this exercise, Interim Provost John Lach said.

"I have worked closely with Professor Szajnfarber since I joined GW in 2019 as dean of GW Engineering, and she has continually made an impact as an outstanding leader with a remarkable aptitude for problem-solving and for making complex issues manageable," Lach said. "I am thrilled that she will be dedicating her considerable talents to leading this important effort on behalf of the university."

As powerful AI tools have become available to the general public in the last few years, GW has taken a position as a leader in the field of trustworthy AI. The university has been awarded two major grants: The multi-university, research-focused Institute for Trustworthy AI in Law and Society (TRAILS), funded by a $20 million grant from the National Institute of Standards and Technology (NIST) and National Science Foundation (NSF) and led at GW by David Broniatowskiand Susan Aaronson; and the $3 million NSF Research Traineeship on the Co-Design of Trustworthy AI in Systems (DTAIS), focused on doctoral education.

Building on that foundation, GW TAI was launched in 2024as a way to build an interdisciplinary community across campus. In its first year, researchers representing all 10 GW schools have affiliated with GW TAI, working together to deliver innovative educational programs, drive interdisciplinary research and inform AI policy with peer-reviewed evidence.

GW's trustworthy AI efforts are focused on understanding AI-infused systems in their context of use, Szajnfarber said, linking technical choices during design with practicable governance strategies. That emphasis, combined with GW's many robust partnerships with research and policy organizations, gives the school an edge in terms of establishing its own strategy-insight into "questions of the day" around AI, a high-level view of what works and what doesn't (and for whom), and how to approach the technological transition on both an institutional and a human level.

"Many AI institutes are leaning into either algorithm development and computing or into a focus on values and ethics," Szajnfarber said. "We're also working on those important topics, but our collective strength as an institution is in the bridge between them. GW's stance between technological innovation, effective governance and policymaking gives it unique insight into questions at the intersection of technology and policy."

Szajnfarber's background spans the disciplines. As a high school student she received a diploma in fine arts, choosing painting with a sculptural aspect as her primary medium. ("There was a lot of chicken wire involved," she recalls with a laugh.) Her undergraduate major at the University of Toronto was in engineering science, a field in which students have to spend the first two years studying every major engineering discipline before choosing a specialization.

As a graduate student at the Massachusetts Institute of Technology, Szajnfarber paired an aerospace engineering master's degree with a second in technology policy, finishing with a Ph.D. in engineering systems. She decided to join GW in part because it gave her the opportunity for appointments in multiple fields; she currently holds secondary appointments in computer science at GW Engineering and in space policy at the Elliott School of International Affairs, with the opportunity for close collaboration with industry and government.

Szajnfarber looks forward to the opportunity the mapping exercise presents to immerse herself in different disciplines' approaches to AI's pressing questions, she said.

"One of the things I'm excited to learn about is all the really interesting ways people are pushing the frontier within their disciplines," she said. "I'm looking forward to seeing how much we can learn from each other and support each other a little bit better, whether that's in research, education or operations."

Szajnfarber tracks her "formal" immersion in trustworthy AI to 2021, when she and GW colleagues Rachelle Heller, Robert Plessand Ekundayo Shittureceived the DTAIS grant designed to train the next generation of AI leaders and researchers. But her involvement with the technology traces back to her time as an undergraduate, when she co-founded and led the University of Toronto's artificial intelligence and robotics club. (Their achievements included placing second in an international robot firefighting competition and "almost" building a team of soccer-playing robots, she said.) When Szajnfarber moved into her professional sphere, aerospace engineering, she entered a field that had been incorporating AI into safety-critical systems for decades. "So I have been around human-AI collaboration for most of my career."

And as a systems engineer, Szajnfarber said, she's comfortable breaking down and analyzing the operation of large and complex organizations like GW. "I'm part systems engineer, part organization theorist-and I'm also an empirical researcher, so I appreciate the value of this work as a data set as well," she said. "I think it's a good example of GW leveraging the strengths of their faculty for other things that are important."

The mapping exercise's three areas of focus provide scaffolding for its analytical goals:

• Research
  • Understand AI as an enabler (and complicator) of the research enterprise, in terms of both research itself and research administration
  • Provide researchers and staff administrators the tools to engage productively and responsibly with these technologies
• Education
  • Understand what GW students need to know about AI and what skills they need to thrive in a post-graduation world where AI will be commonplace
  • Understand how the presence of AI in the classroom should change teaching and learning
  • Determine what supports are necessary for the people at the center of that evolution, including tools, training and beyond, and provide them
• Operations
  • Grasp the opportunities and risks of using AI as part of workflows and the complexities of moving in that direction, including cost and energy use
  • Determine and provide necessary supports for staff

Szajnfarber said she hopes the information collected by the mapping exercise will serve not only as a foundation for the university's AI strategy, but also as a powerful resource for community members. Faculty, for instance, will be able to share creative and meaningful ways they've implemented AI in research and teaching-something already happening across GW, from Alexa Alice Joubin's approach to English literatureto John Paul Helveston and Ryan Watkins' work on the uses of large language models (LLMs) in scientific research. They'll also be able to explore how their colleagues help students build skills for information literacy and a rigorous, realistic approach to these technologies.

Both sets of skills will be crucial to equip students coming of age in this evolving technological landscape, Szajnfarber said.

"As a university, we have the unique opportunity to give students a chance to explore in a safe environment-to play, ask questions and sometimes fail, while being guided by experts in their fields," she said. "As the whole world wrestles with how work will change in the age of AI, as a GW community we're well positioned to grow together and chart the course."

Leading the Way to AI You Can Trust