On behalf of the SFPE Foundation, we are thrilled to reflect on the incredible energy, collaboration, and innovation that defined the 2025 AI in Fire Engineering Summit. Held from May 28-30, 2025, in Berkeley, California, the event brought together a diverse and passionate group of over 160 attendees from more than 19 countries and 20 U.S. states. This landmark event, a key part of our Grand Challenges Initiative (GCI), marked a pivotal moment for the future of fire engineering.
The summit's core mission was to explore the transformative potential of artificial intelligence in our field. As our world faces new technologies, systemic challenges, and the impacts of climate change, it is critical that fire engineers are equipped with the most advanced tools and knowledge. This summit was designed to bridge gaps, showcase current and future AI applications, and, most importantly, identify the key areas where our collective efforts can have the greatest impact.
What Did We Talk About?
The summit kicked off with insightful opening remarks from Dr. Leslie Marshall of the SFPE Foundation, Dr. Michael Gollner of UC Berkeley, and Dr. Tsu-Jae King Liu, Dean of the College of Engineering at UC Berkeley.
Our opening plenary session set a powerful tone for the days to follow.
- Dr. Ilkay Altintaş, Chief Data Science Officer at the San Diego Supercomputer Center, delivered a compelling keynote on "Demystifying AI for Fire Science and Management". She emphasized the need for actionable science solutions that deliver the "Right Model and Right Data at the Right Time" and highlighted the importance of collaborative platforms like the Wildfire Commons to accelerate innovation.
- Dr. Matthias Ihme of Stanford University followed with "Artificial Intelligence as an Agent to Suppress Fires," exploring how machine learning can address challenges in wildfire management. He traced the history of ML, acknowledging challenges like data availability and the "black-box" nature of some models, while pointing toward the promise of physics-informed machine learning (PIML) to ensure more robust and physically consistent results.
A dynamic panel discussion, "What is AI, and where is it heading?" further explored the integration of AI into our profession, covering everything from digitalization and ethics to the future role of the fire protection engineer and the need for trustworthy AI education frameworks.
The summit's technical sessions were structured around the four key topic areas identified in the GCI's Digitalization, AI, & Cybersecurity white paper. Read on to learn more about what each session covered and examples of speakers.
- Data Management & Analysis for Risk Assessment
- Presentations in this area highlighted the foundational importance of data. Dr. Ali Tohidi from the University of Maryland discussed the critical work of curating synthetic and high-resolution observational data for wildfire studies, while Dr. Shuna Ni, also from the University of Maryland, presented on automated fire pattern recognition to bring more objectivity to fire investigation.
- Smart Fire Modeling
- This section showcased the cutting edge of predictive modeling. Dr. Michael Gollner of UC Berkeley presented research isolating the primary drivers of fire risk to structures in California. Dr. Khalid Mosalam, also from UC Berkeley, introduced a novel approach to predicting the "Most Fire-Sensitive Point" in building structures, drastically reducing the computational expense of safety verification. Dr. Qingsheng Wang of Texas A&M shared his work on developing machine learning models to predict flammability and flammable dispersion.
- Digitalized Design, Inspection, Operation & Maintenance (DIOM) of Fire Protection (FP) Systems
- The focus here was on the practical application of AI in the lifecycle of fire protection systems. Dr. Jonathan Hodges of Jensen Hughes addressed the essential topic of verification and validation (V&V) for AI systems used in DIOM, while David Flack of Cardinal Health provided a compelling industry perspective on using AI for property loss control in warehouses, including monitoring housekeeping and detecting smoke with existing camera systems.
- Digitalized Firefighting and Evacuation
- This final set of talks explored how AI can directly support first responders and improve life safety. Dr. Xinyan Huang from The Hong Kong Polytechnic University demonstrated AI applications in smart firefighting, including smart sprinkler design and real-time heat release rate estimation from video feeds. Dr. M. Hamed Mozaffari of the National Research Council Canada presented groundbreaking work on using computer vision and AI for the early detection and prediction of flashover, a critical tool for firefighter safety.
What Were the Workshop Outcomes?
The summit culminated in a collaborative workshop to define our path forward. Through breakout sessions and a final prioritization vote, attendees identified the most pressing needs for the fire engineering community.
Top Research Priorities:
- Data, Data, Data: The highest-voted priority was a combination of needs related to Data Accessibility, Standardization, and Utilization. This includes creating frameworks for data sharing, defining minimum data requirements for FPEs, and breaking down data silos.
- Creating Authoritative AI Model Standards: The second major priority is the development of standards for AI models, with a strong emphasis on reproducibility, uncertainty management, quality assurance, and V&V.
- Improving Building Design and Risk Assessment: There is a clear demand for leveraging AI to improve the process of building design and to develop more robust, science-based risk assessment methods.
Top Education & Outreach Priorities:
- Developing Educational Materials and Curricula: The top priority was the creation of educational materials that leverage AI for improved QA/QC practices and staff training, along with workshops on dynamic simulation techniques.
- Integrating AI Early in Design: A strong consensus emerged around the need for training programs that integrate AI into the fire safety design process from the very beginning.
- Fostering a Collaborative Community: Attendees emphasized the importance of continued conversation between experts, FPEs, academia, the fire service, and regulators to ensure fire protection is not an afterthought.
What Next?
The 2025 AI in Fire Engineering Summit was more than just a conference; it was a call to action. The SFPE Foundation is committed to acting on these priorities. We will be funding projects emerging from this summit and have formed an AI in Fire Engineering Task Group within the GCI to continue this vital collaboration. Interested in serving on the task group? Get in touch.
We extend our deepest gratitude to our hosting partners—the NFPA Research Foundation, UC Berkeley Department of Mechanical Engineering, and the PEER Center—and to our Program Committee, co-chaired by Amanda Kimball, Dr. Leslie Marshall, and Dr. Michael Gollner. We also thank our generous sponsors for making this event possible:
- Visionary Sponsors: Firescore and GAe Engineering.
- Contributor Sponsors: Code Red Consultants, Fire Safety Research Institute (FSRI), LMDG, and Telgian Engineering & Consulting.
Thank you to everyone who participated. The path forward is clear, and by working together, we can harness the power of AI to build a safer world.