Guardrails removed spam, off-topic, unclear, or duplicate replies.
Ask a question below.
Published answers will appear here.
Thanks for the question Glen! We had much more on this in an older version but ended up letting it drop out, which I think was a mistake. No, I don’t think such groups can possess the two moral powers—I think they are rendered as persons as a legal fiction, for the sake of achieving specific instrumental goals (which is also an approach we could take with AIs of course). I think corporations are reducible to the natural persons who comprise them, plus some institutional structure. The corp itself can’t be rationally autonomous (revise, form and pursue a conception of the good), any appearance of such is actually its constituent members’ autonomy; similarly for moral competence.
@sethlazar @nedhw @sethlazar I may have missed it, but I don't see in your piece any discussion of the potential personhood of "corporate" entities, such as nations, cultural groups, corporations, etc. Do you believe such groups can satisfy your definition? If not, why do you think machines can?
@sethlazar @nedhw But you think AI systems are more like having these things? This is where I think you and I would diverge. I am more bullish on these things for “corporate” entities and relatively less so for AIs. @m_t_prewitt would be a good person to comment here.
@glenweyl @nedhw Oh I also missed the other potential group agents you mentioned; I’m much more sceptical that cultural groups and nations constitute agents of any kind; for states the reply is the same as corps.
@sethlazar @glenweyl @nedhw So if one could mathematically reduce an AI model to the people that comprise them (namely their writings), would that change your perspective on AI personhood?
@sethlazar @nedhw Or at least not in the Kant Rawls sense. I also think AI personhood has far more risks of double counting
@sethlazar @nedhw I also think in many cases collective entities are closer to satisfying the PCP than are individual humans
@sethlazar @nedhw @m_t_prewitt Overall I don’t see a very meaningful distinction between corporate entities and AIs
Guardrails removed spam, off-topic, unclear, or duplicate replies.
Ask a question below.
Published answers will appear here.
@sethlazar @nedhw Or at least not in the Kant Rawls sense. I also think AI personhood has far more risks of double counting