Current Projects

Our current research spans three interconnected questions: how people and institutions think and decide alongside AI, how technology can be co-designed with the people who use it, and how public sector systems can adopt AI responsibly and effectively. These projects study emergency managers, public administrators, and planning professionals, using methods that range from participatory co-design workshops to neuroimaging and psychometric measurement. The lab's environmental governance work extends these same convergence and stakeholder-engaged methods to freshwater resource management.

Current Projects:

AI, Cognition, and Public Sector Work

Public Interest Technology in emergency management: With DC area emergency managers, we co-design a socio-technical system that strengthens emergency managerial competence and autonomy. Our research integrates cognitive, organizational, and societal perspectives on technology use in crisis management. We examine the cognitive correlates of decision-making among emergency management professionals in high-risk scenarios and in scenarios involving AI, to understand how cognitive, dispositional, and motivational factors link to decision outcomes. The lab maintains longitudinal data tracking US public and emergency managerial attitudes toward AI and their cognitive impacts over time.

Designing AI for Judgment: Cognitive offloading in public sector decision-making: This project builds and tests a custom AI tool that prompts reflection and questioning during crisis decision-making, countering the tendency to offload cognitive work onto AI systems. Socratic Sage asks questions instead of supplying answers, encouraging emergency managers to reason through problems actively. The project uses a controlled experimental design, including fNIRS neuroimaging, to measure the tool's cognitive effects, and a Community Advisory Council keeps the design grounded in practitioner needs.

AI and the Future of Public Sector Work: This project uses knowledge co-production and participatory co-design to understand how public managers perceive AI's implications for their work, and which tasks in their routine workflow they consider appropriate or inappropriate for AI integration.

Thinking Dispositions and AI Adoption: This project examines the cognitive, attitudinal, and motivational factors that predict AI adoption in public sector contexts. AI literacy predicts AI adoption, but little is known about the thinking dispositions that predict AI literacy.

Perceived Information Overload and the Changing Digital Landscape: As digital information and communication proliferate with technological advances, we ask whether perceptions of information overload are rising, and what the causes and implications could be. The lab maintains a longitudinal dataset tracking information overload and its impacts over time.

Environmental Governance

Freshwater Salinization Governance: This project studies the governance of freshwater salinization in the Occoquan Watershed, combining Ostrom's social-ecological systems framework with stakeholder-engaged convergence science. Working with Stanley B. Grant and Megan A. Rippy, we propose a three-stage Theory of Change for managing salinization and engage stakeholders directly in shaping governance solutions. The project draws on a broader NSF-funded team science effort studying freshwater salinization, and examines how transdisciplinary research teams build the capacity to tackle complex environmental problems.

Past Projects: 

AI Ethics discourse in the public sector: This project developed a picture of artificial intelligence use and discourse in the public sector to guide theory development and identify areas for technical assistance. We analyzed use cases of artificial intelligence (AI) globally and across levels of government and a variety of public sectors (from policing and transportation to social services and public health) to examine the discourse around AI ethics in the public sector. Our analysis was informed by Epistemic Frame Theory, and Epistemic Network Analysis (ENA) was used as the analytical approach. We sought to answer questions such as: What AI ethics concerns were most often discussed in use cases across sectors, countries, applications, and levels of government? Which AI ethics concerns were neglected? What were the key differences in AI ethics discourse across sectors, countries, applications, and levels of government? What were the linkages among AI ethics concepts, and where were linkages missing?

Past Project Events:

August 12, 2024: AI in Hazard Mitigation and Emergency Preparedness Workshop: The PI-Tech Lab event on AI in Hazard Mitigation and Emergency Preparedness drew 30 emergency managers to the EBC at the VTRC in Arlington for a day long workshop on how to better use AI for emergency and crisis management, specifically hazard mitigation planning. The group worked through a participatory co-design process to develop a framework for new AI tools and operational and ethical guardrails and guidelines for using those tools. The group also offered feedback on a pilot custom GPT created by Virginia Tech called Hazard Helper, intended to provide feedback on hazard mitigation plans. The workshop was funded by the Destination Areas 2.0 initiative (Office of the Executive Vice President and Provost) on Public Interest Technology to design ethical AI tools and deploy them responsibly. Click here to access the workshop summary