The research team works alongside Virginia’s Planning District Commissions and the Department of Health to ensure their digital tools meet the practical needs of public safety officials. This strategic partnership addresses the reality that modern utility management has shifted from manual valves to sophisticated internet-based remote monitoring systems. While these advancements increase operational efficiency, they simultaneously expand the digital attack surface of critical national infrastructure. Historically, water treatment plants operated on isolated hardware, effectively creating an air gap between sensitive controls and the public internet. However, the push for connectivity has bridged this gap, often exposing aging legacy systems to global threats. With a $170,000 grant from the university’s Applied Research & Innovation Initiative, researchers are pioneering a multidisciplinary approach to shield this vital liquid resource from contamination or total service disruption.
Predictive Defense Through Innovative Technology
The project addresses a growing security gap created by the integration of aging legacy systems with modern internet-based remote monitoring tools. Many municipal water systems currently suffer from systemic vulnerabilities, such as outdated software and weak password protocols, which are often left unaddressed due to limited local funding or technical expertise. Unlike typical data breaches involving financial information or personal identity theft, a successful cyberattack on water infrastructure can cause immediate physical harm to a population. Potential consequences include the unauthorized opening of dam gates, the contamination of drinking supplies via the manipulation of chemical treatment levels, or the total disruption of water flow to entire communities. By moving beyond standard reactive defenses, the research initiative aims to create a proactive, intelligence-driven shield that considers the physical impacts of digital failures on public health and local ecosystems alike.
PROPHETESS: A New Standard in Cyber Intelligence
The centerpiece of this investigative effort is the development of PROPHETESS, a multimodal agentic framework that seeks to revolutionize how utilities perceive and manage risk. Unlike traditional automated penetration testing that typically generates dense, technical reports that are difficult for non-specialists to interpret, PROPHETESS synthesizes a vast array of diverse data sources to provide a clearer picture. This system incorporates cyber event datasets, real-time news feeds, official security advisories, and even high-resolution geospatial imagery to build a comprehensive threat profile. By aggregating these disparate information streams, the framework can identify historical patterns of digital aggression and predict potential future targets with unprecedented accuracy. This holistic view allows operators to move beyond simple code analysis and understand the geopolitical and social factors that might make their specific facility a target for advanced persistent threats.
Bridging the Gap With Large Language Models
A core innovation of the PROPHETESS framework is its sophisticated integration of Large Language Models (LLMs) to bridge the communication gap between technical analysts and municipal leaders. One of the greatest hurdles in infrastructure security is the translation of complex software vulnerabilities into terms that justify significant financial investment or policy changes. The model component of the system analyzes technical scan results and converts them into clear, actionable narratives that describe the stakes in plain English. For example, instead of reporting a specific buffer overflow vulnerability, the tool can explain how that specific flaw could allow an intruder to shut down water pumps serving a hospital or a residential neighborhood. This level of clarity is essential for plant operators and city council members who must make rapid decisions during a crisis or plan long-term infrastructure improvements without possessing a deep background in computer science.
Collaboration and Real-World Application
Building on these technological foundations, the research team emphasized the necessity of integrating environmental science with cybersecurity to protect vulnerable coastal regions. This approach recognizes that water infrastructure does not exist in a vacuum but is deeply intertwined with the natural landscape and regional hydrology. By collaborating with high-level data scientists and regional experts, the initiative ensures that predictive models account for the unique geographical sensitivities of areas where infrastructure failure could trigger secondary environmental disasters. This cross-disciplinary strategy allows the AI tools to be sensitive to the specific ecological risks associated with infrastructure failure in coastal zones, such as saltwater intrusion or wetland contamination. Furthermore, by grounding digital defense in local geography, the project fosters a sense of shared responsibility among environmental managers, utility operators, and cybersecurity experts who all have a stake in regional safety.
Integrating Environmental and Local Expertise
Grounding the technology in local geography also allows the AI to consider the specific ecological risks associated with infrastructure failure. In many coastal zones, the discharge of contaminated water or the sudden change in water levels can devastate local fisheries and sensitive wetlands. The research team uses this ecological context to refine the risk scores generated by the AI, ensuring that facilities near protected habitats receive the appropriate level of monitoring and protection. This methodology highlights the importance of multidisciplinary research where computer science, biology, and environmental policy intersect to solve modern problems. By viewing water systems as part of a larger, living ecosystem, the project offers a holistic strategy for protecting both human populations and the natural environment. This integrated approach ensures that the digital defenses built today are robust enough to withstand the diverse challenges posed by both malicious actors and a changing climate.
Community Outreach and Stakeholder Engagement
The project concluded with a series of successful engagements between the research team and public safety agencies like the Hampton Roads Sanitation District. Through large-scale tabletop exercises such as the “Cyber Fortress” seminar, federal and state stakeholders stress-tested the state’s preparedness against realistic digital threats. These interactions provided the critical feedback necessary to refine PROPHETESS into a comprehensive decision-support tool. Stakeholders recognized that moving from reactive patching to predictive resilience required a fundamental shift in how digital security was funded and managed locally. By implementing the framework, utility providers established a baseline for continuous monitoring that integrated with existing emergency response protocols. This initiative successfully demonstrated that advanced AI could be made accessible for community-wide protection, ensuring that the American water supply remained shielded from increasingly sophisticated global actors.
