Ashwin gave an introduction to the field of AI safety, which studies how to ensure that AI, especially artificial general intelligence and super intelligence, will be safe and trustworthy.
Concrete Problems in AI Safety, by Amodei et al.