PNNL

Samrat (Sam) Chatterjee, a chief data scientist and team leader with the Data Sciences and Machine Intelligence group at Pacific Northwest National Laboratory (PNNL), was the co-author of a workshop report on agentic artificial intelligence (AI) systems published recently by the Center for Security and Emerging Technology (CSET) at Georgetown University.

The workshop report, “Through the Chat Window and Into the Real World: Preparing for AI Agents,” was based on findings of a CSET-led workshop held in May with invited participants from academia, industry, national laboratories, and government. The workshop focused on what constitutes an AI agent, how the technology is improving, what risks agents exacerbate, and intervention points that could help.

“Researchers are actively building AI systems to execute complex tasks in the real world with increasing levels of autonomy,” said Chatterjee. “Recently, significant progress in large language models has further fueled optimism about the prospect of building general-purpose AI agents.”

AI is increasingly critical to PNNL’s mission areas and it is changing how the Laboratory approaches science. PNNL connects AI research and engineering to develop world-class solutions in science, energy, and security problems. PNNL has expertise, a unique location, strong and growing partnerships, and distinguishing capabilities.  The Center for AI @PNNL coordinates PNNL AI expertise to drive advances in AI and its applications with partners, including the Department of Energy, academia, sponsors, other federally funded research centers.

Chatterjee is a chief data scientist and team leader with the Data Sciences and Machine Intelligence group within the Advanced Computing, Mathematics, and Data Division. He leads national security and computational science research efforts in support of multiple Department of Homeland Security, Department of Energy, and Department of Defense sponsors with focus on infrastructure network resilience modeling, security risk and decision analytics, and multi-agent learning and optimization. He has authored two books, six book chapters, and over 100 peer-reviewed publications on the infrastructure, risk science, and AI and machine learning domains, including multiple best paper awards. He also holds an affiliate faculty appointment with Northeastern University in Boston, Massachusetts, serves on the editorial board of the Risk Analysis journal, and is a senior member of the Institute of Electrical and Electronics Engineers.