Tamper resistance, moral intelligence, and self-security will be useful for making artificial intelligence safe for small numbers of people, but will do nothing to protect large populations. This is because all forms of self-security and AI safety..
Relevance: 11.911971