SciFM26 at the University of Chicago

SciFM 2026 is organized around three interlocking questions:

1.     Can foundation models be grounded in physical law — and what does 'grounded' actually mean for a scientific model?

2.     What are the specific, unsolved problems in aerospace, energy, and manufacturing that a physics-aware AI could address — and why hasn't the AI industry addressed them yet?

3.     What does it look like when AI participates in scientific discovery at facility scale, in real time, with appropriate uncertainty and human oversight?

Each session is designed to create productive tension between academic and industry perspectives, and between optimistic capability claims and honest assessments of current limitations. Day 1 features four sector-specific challenge panels spanning Aerospace, Oil & Gas / Energy Transition, Advanced Manufacturing, and Biotech & Life Sciences — each structured around the question: what are the problems current AI is not solving, and what would a scientific foundation model need…