GW Researchers Receive New Grants to Study Applications of Trustworthy AI
Students work in the Science and Engineering Hall at George Washington University. File Photo from The GW Hatchet.
Four GW professors are the recipients of grants from a multi-university initiative to research and improve the trustworthiness and societally beneficial implementation of artificial intelligence.
The Institute for Trustworthy AI in Law and Society (TRAILS), which launched in May 2023 with a $20 million award from the National Science Foundation and the National Institute of Standards and Technology, awarded five new grants totaling $515,000 to fund research into how institutions use AI and build trust over time. The grants are helping to fund new projects to measure and understand trust in AI and to improve its use in education, autonomous drones and local government.
The new grants issued by TRAILS — made up of scholars from the University of Maryland, GW, Morgan State University and Cornell University—range from $40,000 to $160,000 and fund projects across disciplines including computer science, engineering, law, public policy and the social sciences. Previously, TRAILS funded eight grants totaling $1.5 million in 2024 and seven totaling $750,000 in 2025.
David Broniatowski, TRAILS’ deputy director and GW professor of engineering management and systems engineering, said successful projects have helped TRAILS identify research areas that necessitate continued investment, though he added that not all TRAILS-funded projects are research seed projects. He said TRAILS supports a range of activities, including research seed projects and broader impacts seed projects.
“As a result, the change in seed funding from year to year reflects a gradual transition from exploring new ideas through seed projects to making sustained investments in areas and projects that have demonstrated promise and success,” Broniatowski said in an email. “Seed funding remains an important way for TRAILS to identify and develop new research directions.”
The new funding follows officials’ April announcement that the University is backing the Trustworthy AI Initiative as its first major research-focused project under the strategic frameworkGW. Broniatowski said TRAILS is “deeply aligned” with GW’s strategic framework, which identifies AI as a pressing societal challenge and calls on the University to combine its interdisciplinary strengths.
“That is precisely the model TRAILS represents: research across computer science, engineering, law, international affairs, public health, education, psychology and information science, connected to government, standards organizations, industry, educators and communities,” Broniatowski said.
Joseph Dunne, a researcher with the GW Institute for Public Policy and a grant recipient, said his work will compare how local governments in D.C. and Maryland implement AI to improve efficiency while protecting data and public trust.
“Somehow, the public sector hasn’t been able to keep up with AI, and that’s the area that we’re sort of zooming in on or focusing on because we want to help,” Dunne said. “We think that our research will produce lessons that will be useful for the public sector or for governments everywhere.”
Dunne, who is working alongside two UMD researchers, Katie Shilton and Jwan Khisro, said the team secured an agreement with the D.C. government to participate in the study before their research proposal was approved. He said the D.C. government is “very interested” in AI research to improve government efficiency and welcomed the project.
“There are obvious efficiency gains coming from AI, but at the same time, there are special challenges like confidentiality of data, cyber risk, and so on,” Dunne said. “Somehow, the public sector hasn’t been able to keep up with AI, and that’s the area that we’re sort of zooming in on or focusing on because we want to help.”
Nan Wu, an assistant professor of electrical and computer engineering in the School of Engineering & Applied Science, said she received a grant with fellow SEAS professor Peng Wei to develop lightweight, robust vision-language models for safe autonomous drone delivery. She said her goal is to make advanced AI reasoning “practical and safe” on resource-constrained drones by addressing potential risks, such as delayed or unreliable decisions caused by environmental obstacles.
Wu said her research shows that trustworthiness is particularly important for AI-enabled physical systems, such as drones. She said for chatbots, a delayed or unreliable response may be “inconvenient,” but for autonomous drones, it can become a physical safety issue.
“TRAILS support enables us to develop the algorithms, hardware-software systems and high-fidelity simulation environment needed to systematically study and mitigate these safety risks,” Wu said.
A University spokesperson said a TRAILS grant is also funding a project Broniatowski, who also serves as GW’s site lead, and Valerie Reyna of Cornell’s College of Human Ecology are leading on how trust can be rebuilt after AI failures.
—Story by Sarah Gross, assistant news editor for The GW Hatchet
