Human–AI Agent Collaboration in Structural Engineering and Code Development
Note: This presentation will not be recorded.
AI agents are large language model systems that plan multistep tasks and execute them using external tools. In structural engineering, these systems can now complete analyses that previously required days/weeks of specialist effort, such as deriving design expressions or computing collapse loads.
This talk examines what these capabilities mean for tackling long-standing problems and for the development and maintenance of building codes. The talk first reviews how agents are being used in structural engineering design workflows. Then, we present case studies from our group in which a lead AI agent orchestrates subordinate agents to derive and verify solutions to structural mechanics problems, while an independent agent audits the resulting claims. These cases include simulation-free analyses of RC columns under fire conditions and a mechanics-based model for the shear strength of steel- and FRP-strengthened beams. Building on these results, the talk outlines how AI agents could support code committees. The talk also covers current limitations of agents in this setting, including fabricated references and unverified numerical output. Accountability and licensure questions are addressed with attention to how agent-generated evidence could enter a consensus standards process. The talk closes with a proposed framework in which agents generate and verify technical evidence while engineers and committees retain responsibility for adoption.
Learning Objectives
Upon completion of this session, participants will be able to:
1. Describe how AI agents plan and execute multistep structural engineering tasks, including the roles of lead, subordinate, and independent audit agents in a verification workflow.
2. Evaluate the reliability of agent-generated structural analyses by distinguishing certified two-sided bounds from unverified numerical estimates and by identifying common failure modes such as fabricated references.
3. Explain how AI agents can support building code development by tracing the technical basis of existing provisions and checking proposed changes against test data, while engineers and code committees retain responsibility for adoption.
1 hr of CE Credit
CE certificates are emailed out one week after the presentation to give time to verify attendance time.
Registration
This is a virtual event on Zoom. All registrants MUST have a free Zoom account to access the webinar and must register with Zoom to receive the link.
Members - $45
SEA Members - $45
Nonmembers - $65
Event registration closes 30 minutes before the webinar starts.
About the Speaker
M.Z. Naser
Associate Professor, Clemson University
M.Z. Naser is an associate professor at the School of Civil and Environmental Engineering and Earth Sciences & a member of the Artificial Intelligence Research Institute for Science and Engineering (AIRISE) at Clemson University. At the moment, my research group is creating Causal & eXplainable machine learning methodologies to discover new knowledge hidden within systems belonging to the domains of Structural engineering and Materials science to help us realize functional, sustainable, and resilient infrastructure. Much of my current projects cover the areas of structural & fire engineering, tailoring properties of construction materials, and retrofitting of aging structures. I am humbled to serve as the chair of the ASCE Advances in Technology committee and be among the top 2% of highly cited scientists worldwide, according to the Elsevier-Stanford study (since 2022 - now).
