In you're in favor of "AI safety" (broad definition), what's your most compelling cast-length argument?

"AI safety" is mostly a distraction from the real dangers of human bad actors. People and people organized as corporations are more than capable of doing harm at scale with whatever technology is available.

That’s part of AI safety 

How do we prevent bad actors from using it to harm people?

- assumes that we’ve solved the alignment problem (very hard) 

- we manage to program some sort of moral compass in the AGI