Depends on your definition of AI Safety. Like you said, there is in fact some very legit research that (unfortunately) is done under the umbrella of "AI Safety", but that's quite different from the threat-to-humanity AI safety crowd.
No comments yet.