Date and time: Thursday 26 March 2026, 13:00-14:00 CET
Speaker: Reza Allahvirdizade, KTH
Title: Built Environment Safety in AI Era
Where: Digital Futures hub, Osquars Backe 5, floor 2 at KTH main campus OR Zoom
Directions: https://www.digitalfutures.kth.se/contact/how-to-get-here/
OR
Host: Smriti Sharma sharma2@kth.se
Bio: Reza Allahvirdizadeh is a postdoctoral researcher at Division of Structural Engineering and Bridges at KTH. His research focuses on structural safety, reliability theory, and computational methods for probabilistic assessment. He received his PhD from KTH and has held research positions at the University of Minho and the University of Tehran.
He has also collaborated with several international research groups, including the Engineering Risk Analysis Group at the Technical University of Munich, the Department of Mechanics of Continuous Media and Theory of Structures at the Technical University of Valencia, and the Structural Engineering Laboratory at the Swiss Federal Laboratories for Materials Science and Technology (EMPA).
Abstract: A functioning built environment has always been a key factor in societal resilience and development. We often interpret the linguistic concept of proper functionality as “safety.” But what is safety, especially in measurable terms? And how have we assured ourselves that our designs are safe throughout history?
In this talk, I will discuss how the mindset regarding this concept and its implementation has evolved from experience-based approaches to modern probabilistic frameworks. This evolution has revealed a critical cognitive gapconcerning the human inability to intuitively sense the very small numbers associated with failure probabilities, and the massive computational resources required to calculate them accurately. I will review the attempts made over the last couple of decades to address this matter, highlighting their shortcomings despite their significant contribution to shaping our modern built environment.
The great computational capabilities of AI have opened new possibilities to overcome these limitations. I will demonstrate how AI-based modeling can transform static post-design checks into dynamic optimizations and real-time decision making to extend the service time of our infrastructure. At the same time, this shift introduces new questions concerning trust and epistemic uncertainties, especially in safety-critical systems. I will highlight these challenges as they outline important research directions for integrating AI into the safety management of the built environment.
