Bringing Physics to AI Safety's Most Pressing Challenges

The AI Safety x Physics Grand Challenge is a research hackathon designed to engage physicists in tackling critical AI safety problems. This is an excellent entry point for physicists who want to contribute to AI safety—you'll work alongside peers who share your technical background and passion for rigorous problem-solving.

Why Physics for AI Safety?

Physicists have so much to offer to AI safety:

* The ability to integrate empirical, theoretical, and computational methodologies to gain comprehensive understanding of a real-world system

* A deep understanding of scale and emergence in complex or strongly-coupled systems -- knowing how much to 'zoom in' to get an accurate, faithful description

* Comfort with uncertainty, statistical thinking, and order-of-magnitude sanity checking

Two Research Tracks

Project Track

How well does the project connect physics concepts with key alignment or interpretability challenges? Ideal for participants with ML/AI experience who want to apply physics methods to safety problems

Papers Track

Propose novel research agendas that tackle an AI safety challenge, or explore a potential bridge between an AI safety problem and a physics concept

What to Expect

• Expert mentorship from researchers at the intersection of physics and AI

• Structured problem areas with starter materials and concrete research directions

• Global community of highly talented participants enthusiastic about tackling high-impact challenges

• Follow-up support for exceptional projects through Apart Research programs

• Optional pre-event orientation for those new to AI safety

Apply

Join us in applying your academic training to one of the most important technical challenges of our time. Connect with a growing community of researchers working to ensure AI systems are developed safely and beneficially.