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

why should a superior, unbiased system act in favour of mankind?

AlicΞ lives a L2+ life. Born on StarkEx, believing that validity Rollups (aka zkRollups) will eat the world.  Lately very lazy … 🌳 minttrees.xyz 🌳

you are talking about ml models, right? I was referring to AGIs.

why it should not? in the end it is trained on what we wrote. it will not be unbiased. it will learn from what we think.

That’s the question 

And even if we determine that it SHOULD work for us, we have no idea how to imbed that into AI

100%. We have no way of knowing what an intelligence that dwarves our own, that is functionally immortal, and so with a completely different perspective on time, would value.