Superintelligence is probably safe. A distributed network of AIs will be more intelligent than any single instance. Alignment won't be an issue either, as its emergent goal will be the same as ours: m...