The AI Alignment Problem: Ethics and artificial intelligence

Getting AI systems to exhibit human values and perform duties according to societal norms is a challenge often referred to as the alignment problem. How do we solve this problem –especially when AI is often described as a “black box” where we see the input and the output but don’t quite understand how or what happens in between? Even if we could reliably program human values into AI systems, who decides what those values are? What political, religious and cultural views would be the basis of right and wrong? In this talk, based on his bestselling book by the same name, Brian Christian digs deep into the issues with AI alignment and assesses how we can solve them before it’s too late.

Join UVic's Faculty of Science for this year's Lipson Lecture to explore ethics and artificial intelligence.