Long Foreseen, the Problem of AI Alignment Is Finally Reality. Solving It Won’t Be Easy. - SingularityHub

The challenge of making sure AI systems do what humans actually want them to do is now a real, pressing problem, not just a theoretical worry. Getting this wrong could cause serious harm at scale.