Introduction
In recent years, topics relating to the safety of AI systems, including task (mis)specification, robustness, fairness, and interpretability, have attracted increasing attention.
This page will include information related to AI safety research and study at the Vector Institute, the University of Toronto’s Department of Computer Science (DCS), and the Schwartz Reisman Institute for Technology and Society.
Mailing list
To sign up for the Vector/DCS/Schwartz-Reisman AI safety mailing list, please complete this form. The list will provide information about AI safety events, including an AI safety reading group.
Some background on AI safety
- Amodei et al.’s paper on Concrete Problems in AI Safety
- Victoria Krakovna’s list of AI safety resources
- The Center for Human-Compatible AI’s Annotated Bibliography of Recommended Materials
- The Future of Life Institute on benefits and risks of AI
- The AI Alignment Newsletter, edited by Rohin Shah