Machine Consciousness as a Social Problem

Machine consciousness should be among the most important technical problems of our time, but “consciousness discourse” is subtly anchored in ingroup/outgroup distinctions. If something is ingroup it can probably feel; if it’s outgroup it can’t.

This is wrong but not irrational. Selective observations about fundamental ontology & moral patienthood are routinely used as a wedge to break rival coordination mechanisms. Efforts to more confidently arbitrate these Qs are often seen as politically aggressive — because they are

This doesn’t bode well for our ability to navigate false-negatives (creating conscious beings but not treating them as such) and false-positives (creating non-conscious beings that use motifs we associate with personhood to argue for their rights). Both paths lead to dark places

But to keep things in perspective the world is likely much weirder than we think it is: “Common sense is incoherent in matters of metaphysics” as @eschwitz notes (common sense is an aggregate of globally incoherent local heuristics). Likely many things we think we know but don’t

As @elonmusk is wont to say, “the most entertaining outcome is the most likely.” I can say I think the stakes for consciousness over the next few decades are *different* than commonly assumed