--- Nathan Barna <[EMAIL PROTECTED]> wrote: > On 9/12/06, Stephen Reed <[EMAIL PROTECTED]> > wrote: > > such. How do imagine a safe upload of all > humanity > > would unfold? > > Other good exploitive persuasion probably'll come > along, but it could > be conventional that each person uploads into her > infinitesimal > polymorphic hypercomputer and then, afterward, when > not being friendly > becomes laughably impossible, externally > communicates according to > volition. Hm. For the sake of your point, I suppose that all humans upload according to your scheme and no bodies remain to be harmed by the AIs. But I wonder how safe separate AI's would be, prior to their voluntary federation? Would it be possible that infinitesimal polymorphic hypercomputers are in fact obscure enough to be unharmed by a malicious AI born of the same technology? Do you imagine a safe exodus from Planet Earth to obscurity and a subsequent voluntary return from the diaspora?
The point of my questions is that I believe that it is ultimately simpler and safer to devise an AI prior to uploading that behaves in a friendly fashion, following human laws and ethics, even if it is smarter than its human developers. The US military will have to address this issue soon with upcoming autonomous weapon systems. Presently their safeguards consist of keeping a human in the control loop. But how is the behavior of a fully autonomous robot controlled via its design? -Steve __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com ----- This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/[EMAIL PROTECTED]
