3 ms·
In my experience, the AI Safety people are less concerned with a nebulous "morality", and more concerned with practical things like making AI self-improve slowl
by derefr 7y ago
In my experience, the AI Safety people are less concerned with a nebulous "morality", and more concerned with practical things like making AI self-improve slowly enough that we can catch it before it's a runaway process; or making AI not think in ways where it imagining torturing people means there's an actual ephemeral consciousness experiencing the qualia of being tortured. Engineering stuff, that happens to interact with morality... in about the same way that genetic engineering could be said to interact with the morality of the resulting organism.