Artificial Intelligence in Engineering: NAE Regional Meeting at UC Berkeley

Keynote Speakers 

The opening keynote session focused on AI’s role in infrastructure. Kenichi Soga, Donald H. McLaughlin Professor in Mineral Engineering and a distinguished professor of civil and environmental engineering at UC Berkeley, discussed how AI can revolutionize the way civil engineers approach aging infrastructure. Soga explained how engineers are designing machine learning models that can detect weak points in systems like urban water pipes, predicting failures before they occur. These models could prevent damage and save thousands of dollars.  

In the second keynote, Tianzhen Hong, senior scientist at Lawrence Berkeley National Laboratory, described how AI is helping make building operations more energy efficient, resilient, and demand flexible. Hong introduced the use of surrogate models that combine engineering and data-driven insights to deliver more accurate predictions across the building lifecycle, from planning to demolition. 

In the afternoon keynote, Tejas N. Narechania, professor of law at UC Berkeley, raised critical concerns about AI’s concentration of power and its implications for democracy, resilience, and equality. He cautioned that many AI models are designed with a preference for self-optimization, creating scenarios where systems could fail if dominated by a single provider that is unable to deliver. This creates tight supply constraints that make critical infrastructure vulnerable. He offered mitigating solutions, including establishing public options for data and cloud services and enacting nondiscrimination rules for AI model owners to ensure fair access and competition. 

In her keynote, Kristin Persson, Daniel M. Tellep Distinguished Professor in Engineering at UC Berkeley, reflected on how engineering data collection has evolved from the time of Thomas Edison to today’s AI-driven landscape. She highlighted the leaps in technological capability over hundreds of years and underscored that AI could streamline these leaps. Persson showcased how AI has enabled the discovery of innovative materials, including an alkaline cathode developed using AI-guided data analysis, and emphasized that we will continue innovation through AI. 

In the closing keynote, Pieter Abbeel, professor in AI and robotics at UC Berkeley, dubbed the current era “the humanoid summer,” describing how AI is driving rapid progress in humanoid robotics. Abbeel highlighted the use of tele-creating methods, where robots learn from human video demonstrations, and encouraged engineers to get involved in using simulation tools, even without physical robots. He noted that there is currently no clear consensus on how best to build AI for humanoids, making this an open frontier ripe for innovation. 

Lightning Talks A series of lightning talks showcased how AI is driving innovation across diverse disciplines, from scientific discovery and electron microscopy to healthcare, robotics, and data analysis. 
Speakers included Aditi Krishnapriyan, Mary Scott, Alane Suhr, Emma Pierson, Dani Ushizima, and Angjoo Kanazawa. The talk ended with a video of the Lawrence Berkeley National Lab. 

Panel Discussion: AI’s Transformative Power in Engineering The meeting closed with a panel discussion featuring Bryan Catanzaro, vice president of applied deep learning research at NVIDIA, and UC Berkeley professors Ion Stoica and Dan Klein. Moderated by Tsu-Jae King Liu, the conversation explored professional perspectives on the future of AI and its implications for engineering disciplines. 

Panelists reflected on the ethical challenges posed by AI, including bias in data, the risks of centralization, and the importance of ensuring AI systems are transparent and trustworthy. They emphasized that as AI continues to accelerate, so too must industry, academia, and policymakers collaborate to ensure its responsible and inclusive deployment. 

The discussion also addressed the pressures of keeping up with AI’s pace of change and the need for cross-sector ecosystems that democratize access to AI tools and resources, particularly for small and mid-sized enterprises navigating rapid technological shifts. 

Looking Ahead 

The NAE Regional Meeting at UC Berkeley’s central theme was that while AI transforms engineering disciplines, its future will depend on the ways we choose to govern, deploy, and scale it. Whether it’s embedding sensors in aging infrastructure, designing ethical frameworks for AI models, or exploring the frontiers of humanoid robotics, the meeting highlighted the need for engineers to lead with curiosity, caution, and a commitment to equity. 

Photos by Adam Lau/Berkeley Engineering