Under the leadership of Ketan Savla of USC’s Viterbi School of Engineering, and assisted by Jun Zhuang of the University of Buffalo and Sam Chatterjee of the Pacific Northwest National Laboratory, CREATE announces a new research initiative: Securing Critical Infrastructure in the Age of AI (SCIAI). SCIAI is an interdisciplinary program drawing from CREATE’s talent in computing, resilient infrastructure, risk assessment and economics, and will include faculty from multiple USC schools along with CREATE’s national and international research partners. For more information, contact CREATE’s Director, Randolph Hall ([email protected]). SCIAI’s focus is described below.
Securing Critical Infrastructure in the Age of AI (SCIAI)
For more than two decades, CREATE has helped the nation understand what threats to critical infrastructure cost in lives and economic activity. Artificial intelligence is now moving into the systems CREATE was established to protect, including electric power, transportation, water, communications, ports, and supply chains. The scale is already visible in energy: the International Energy Agency estimates that widespread use of existing AI applications could save power plants up to $110 billion annually and unlock 175 gigawatts of transmission capacity by 2035. Yet the potential gains remain largely unquantified, as do the consequences of AI failure, AI-enabled attacks, and missed opportunities to use AI defensively. CREATE will develop the evidence, methods, and public-policy frameworks needed to deploy AI securely, strengthen infrastructure resilience against threats, and prevent systemic consequences.
A distinctive contribution: consequence-informed AI security and resilience. CREATE will connect decisions about AI deployment, defense, autonomy, and human oversight to quantitative models of societal and economic consequence. This approach will assess whether AI can perform safely, how adversaries might exploit it, and where AI-enabled defenses can reduce risk. Researchers will use CREATE’s risk-analysis methods to set thresholds for moving from human-directed analysis to AI recommendations and bounded autonomous action, determining what AI may do, how it must be protected, and when human intervention must be retained. This effort will be complemented by CREATE’s leadership in economic consequence and benefit-cost analysis for safety and security investments.
An integrated research agenda. The work will address three interrelated challenges: protecting infrastructure from adversaries who use AI to discover vulnerabilities, evade defenses, manipulate operational data, or coordinate cyber-physical attacks; using AI defensively to detect intrusions, anticipate equipment failures, and support response to wildfire, extreme weather, and other hazards; and ensuring that AI embedded in infrastructure remains reliable, controllable, and resistant to manipulation. Because infrastructure systems are interconnected, CREATE will also examine how an attack, compromised AI agent, or physical disruption in one network can propagate across others and how it can be detected, isolated, and contained. Finally, CREATE will apply its expertise in resilience analysis to identify ways that affected systems can recover more quickly and efficiently to keep losses to a minimum.
From research to national capability. CREATE will turn its research into an independent capability for testing AI before it is entrusted with critical infrastructure. Infrastructure-scale simulations and testbeds will reproduce AI-enabled attacks, equipment failures, extreme-weather events, and cascading disruptions, allowing researchers to determine whether AI-enabled defenses can detect and contain them before essential services are lost. Demonstrations with government and industry, coupled with evaluation, will help operators and regulators decide where AI can be deployed, what authority it should receive, and when human intervention must be retained.
Why CREATE and USC. CREATE offers a rare foundation for this work: it was the Department of Homeland Security’s first university Center of Excellence and has spent more than two decades translating complex technical threats into economic and policy tools for the government, encompassing exceptional students and faculty from throughout the university. CREATE’s foundation is magnified by the Viterbi School’s Information Sciences Institute and its Stevens School of Computing and Artificial Intelligence, in cybersecurity, cyber-physical systems and AI. It is also complemented by the Price School of Public Policy’s expertise in economic modeling, resilience analysis, security, public safety and national defense.
CREATE’s national and international network of scientists and faculty is anchored in Los Angeles, where wildfire, scarce water resources, extreme weather, seismic activity, infrastructure interdependence, and emerging technologies converge. The 2028 Olympic and Paralympic Games create a timely horizon for research on resilience under extraordinary demand, security, and continuity requirements.
A durable component of CREATE’s future. Within CREATE’s broader vision, sustained philanthropy would keep this capability independent and focused on public consequence rather than product cycles or individual grants. By sustaining essential expertise and partnerships, CREATE would help ensure that AI strengthens the critical systems on which society depends.
Click Here for More Information on CREATE’s AI Research
Posted September 17, 2026