November 15, 2024
Staff Accomplishment

Researchers at PNNL Contribute to AI Security Report

Center for Security and Emerging Technology publishes report based on June workshop

DeCillis and Richards

Researchers at PNNL Marisa Decillis and Luke Richards 

(Composite image by Shannon Colson | Pacific Northwest National Laboratory)

Researchers at Pacific Northwest National Laboratory (PNNL), Marisa Decillis and Luke Richards, contributed to a recent Center for Security and Emerging Technology (CSET) report, “Securing Critical Infrastructure in the Age of AI.”

The report, published in October, was based on ideas exchanged at a CSET workshop in June. “As critical infrastructure operators and providers seek to harness the benefits of new artificial intelligence capabilities, they must also manage associated risks from both AI-enabled cyber threats and potential vulnerabilities in deployed AI systems,” says an introduction to the report. “This report synthesizes our findings, drawing on lessons from cybersecurity and insights from critical infrastructure sectors to identify challenges and potential risk mitigations associated with AI adoption.”

DeCillis, an adviser in PNNL’s Cyber Resilience Foundations Group, and Richards, a data scientist at PNNL, served as workshop panelists.

DeCillis was on the “Challenges in Implementing and Securing AI Systems” panel.

“One of the impressive things to emerge from our panel discussions was the challenge of small organizations to implement AI protections within their systems,” said DeCillis, who also serves as deputy leader of PNNL’s Resilience through Data-Driven, Intelligently Designed Control (RD2C) initiative. “A large electrical company is more likely to have the resources to hire new staff with AI expertise.  But for smaller providers, hiring additional staff with AI skills can be challenging—especially if these small companies are in rural areas. They may have to look at upskilling current IT staff.”

DeCillis also participating in providing additional comments and resource material related to NIST’s (National Institute of Standards and Technology’s) publications related to their AI Fundamental Research portfolio. 

Richards served on the “Constructing a Technical Mitigation and Policy Toolkit” panel.

“It was a much broader discussion on what the risks to critical infrastructure are and how AI can both help address and exacerbate these risks,” said Richards, who serves as principal investigator on projects focusing on assured AI efforts in machine learning safety, security, and usability. “This discussion provided details for the recently published report in CNET.”

Richards also participated in shaping and writing the technical details in the CSET report on how risks could emerge when applying AI to systems that involve both information technology and operational technology. “Applying AI to both systems could lead to information leakage,” Richards said.

DeCillis and Richards are part of The Center for AI @PNNL, which is driving a research agenda that explores the foundations and emerging frontiers of AI, combining capability development and application to mission areas in science, security, and energy resilience. The Center for AI @PNNL is partnered with the Department of Energy Frontiers in Artificial Intelligence for Science, Security and Technology, an initiative to harness AI for scientific discovery, address energy challenges, advance national security, attract and build a talented workforce, and develop technical expertise necessary for AI governance. 

Published: November 15, 2024