Join Noah and Marty as they delve deep into the immense conceptual and computational difficulties involved in ensuring advanced AI remains aligned with human ethics and values over the long-term. We discuss using "negotiation games" to help align AI systems' goals with human values and interests over time. The idea is to incentivize AI systems to make accurate long-term forecasts that benefit humanity.

Also we touch on paradoxes and intractable complications that plague the approaches and thinking at the foundations of current deployment of AI systems.