Three faculty members in the University of Tennessee, Knoxville Tickle College of Engineering are co-principal investigators (co-PIs) on research projects funded by the Department of Energy’s new Genesis Mission, a $5 billion program to create a national ecosystem of AI tools and services that will help researchers address pressing energy, scientific, and engineering challenges.
Department of Materials Science and Engineering (MSE) Weston Fulton Professor Sergei Kalinin is a co-PI on AlphaFilm, a project that will create the nation’s first closed-loop agentic AI for semiconductor thin-film material design.
MSE Associate Professor Mahshid Ahmadi is a co-PI on a project that defends agentic AIs in automated scientific workflows from adversarial attacks.
The DOE awarded funding to 277 Phase I proposals from across the country, representing 342 participating institutions. Over the next nine months, the participants will work to bring their novel AI tools, agents, and workflows to life. Projects that achieve promising results will have the opportunity to apply for additional Phase II funding in 2027.
The materials discovery pipeline requires iterating between theory, synthesis, characterization, and feedback in a loop that can take months or even years. Kalinin’s research group at the University of Tennessee has developed a method of high-throughput scanning probe microscopy that accelerates that process by a factor of 30.
“UT is the only place that can accelerate characterization that much,” Kalinin said.
AlphaFilm is being led by Mat3ra, a company that provides cloud-based infrastructure to streamline the discovery and development of new semiconductor thin-film materials. AlphaFilm will combine Kalinin’s uniquely rapid characterization method, Mat3ra’s theoretical and computational workflows, and the outstanding thin-film growth capabilities at the National Laboratory of the Rockies in Colorado into a single closed-loop workflow.
“By combining (our strengths), we can make a meaningful difference in the rate at which new functional materials are discovered and optimized,” said Kalinin.
As AI agents become increasingly involved at every step of scientific discovery, from generating hypotheses to designing and conducting experiments, they also introduce new security risks.
“An adversarial attack or corrupted information could quietly influence an AI agent’s decisions, potentially leading to incorrect scientific conclusions, wasted experiments, or unsafe experimental decisions,” Ahmadi explained.
Ahmadi and Min H. Kao Department of Electrical Engineering and Computer Science Assistant Professor Suya are both co-PIs on a Genesis project led by North Carolina State University Assistant Professor Xiaorui Liu. The team will develop AI systems that can recognize and defend against these evolving threats while maintaining the accuracy and efficiency needed for autonomous science.
Suya’s lab will work to identify different ways AI agents can be attacked, then determine what types of tests will rigorously evaluate the team’s new defenses. Ahmadi will integrate the project’s AI security framework into her self-driving materials laboratory, where AI agents connect hypothesis generation and computational screening with automated synthesis and real-time characterization.
“This strategy allows us to stress-test the cybersecurity approaches in an actual autonomous scientific workflow and ultimately demonstrate that agentic AI can accelerate materials discovery while remaining trustworthy and resilient,” Ahmadi said.