AI Safety Meets Robotics: The Case for Physical AI Safety
Join Cornell AI Alignment (CAIA) for a guest talk and Q&A with Bear Häon and Kaylene Stocking, researchers and co-founders of the Physical AI Safety Institute (PAISI).
As AI advances beyond chatbots and digital assistants toward increasingly autonomous robots capable of acting in the physical world, new safety challenges are emerging. How can we ensure that AI-powered robots behave safely, remain under human control, and reliably pursue their intended goals?
Bear and Kaylene are working to address these challenges through research on robot foundation models, mechanistic interpretability, and AI alignment. They also co-founded PAISI, a nonprofit research organization dedicated to advancing the safety of embodied AI through research, fellowships, workshops, and educational initiatives.
In this talk, we'll explore the emerging field of physical AI safety, the challenges of aligning increasingly capable robotic systems, and promising approaches to making these systems safer and more interpretable.
The talk will be followed by an open Q&A and discussion.
When: Tuesday, October 27, 5:00-6:00 PM ET
Location: TBD
Food will be provided!
Recommended reading: