AI ethics: Working towards building an AI that will not embarrass itself at a mid-tier dinner party in San Francisco.
AI alignment: Advancing capabilities while feeling vaguely guilty about it.
Accountability in AI: Ensuring AI doesn't interfere with our ability to create sinecures.
AI we can trust: Ensuring AI doesn't discredit the linguistics work of Noam Chomsky.
AI policy: A long march through the institutions for the purpose of solving a problem that will kill us in five years.
Decentralized AI: The token is totally necessary and going to the moon.
Since there is often confusion about the difference between "AI ethics" and "AI alignment", I thought I'd make this handy guide.
AI ethics
AI alignment