AI capabilities are advancing quickly.
Systems are becoming more capable across reasoning, science, coding, persuasion, and autonomous action. Safety and governance may not be keeping pace.
As AI systems become more capable, the gap between what they can do and our ability to reliably understand, constrain, and govern them may become one of the most important problems we face. SafeAI@Penn is a student community working to understand those risks and help reduce them.
Our mission
We help students develop the conceptual, technical, and policy background to engage with AI safety, find collaborators, and take their next step into research, governance, or other high-impact work.
Why AI safety?
Three ideas that motivate our work.
Systems are becoming more capable across reasoning, science, coding, persuasion, and autonomous action. Safety and governance may not be keeping pace.
Highly capable systems could amplify biological, cyber, political, and other risks. If we eventually build systems that are more strategically capable than their operators, reliable control may become a central technical problem.
The field needs people from engineering, policy, biology, philosophy, economics, and more. We help Penn students find tractable problems, build skills, and connect with people working on them.
Programs
A guided introduction to the major ideas in AI safety, from alignment and evaluations to governance, biosecurity, and long-term risk.
Apply nowAccessible sessions with researchers, practitioners, and students exploring important questions in AI safety.
Learn moreMeet collaborators, find mentorship, explore projects, and discover research, internship, and career opportunities.
Join usGet involved
No prior AI safety experience is required. Students from every major are encouraged to apply.
Apply now