Quantitative AI Safety Initiative

Research Leads

Stuart Russell
Berkeley

Dawn Song
Berkeley

Max Tegmark
MIT

Yoshua Bengio
Mila/Univ. of Montreal

Steve Omohundro
Independent

Mission: Place AI safety on a quantitative foundation 

Who we are: QAISI (pronounced “Casey”) is a collaboration founded by AI researchers at Berkeley, MIT and the University of Montreal to place AI safety on a solid quantitative foundation. Engineers in other safety-critical fields have successfully quantified failure rates for technologies from jet engines to nuclear reactors and used this understanding to make them much safer. Our mission is to do the same for AI systems that will impact people’s lives.

Please join us! You’ll find a first job posting here, and more coming soon.

Questions? tegmark@mit.edu