I guess one point of contention I would have in this debate is that it seems to be black and white, either they will obtain sentience or they won't, and the implication is that if they do then all will do so. But, why would that be the case? Why wouldn't there be a hierarchy of intelligence capabilities, from basic automation on one end --> batch machine learning --> continuous on-line learning --> higher-level reasoning in AI --> near human-level AI agents used in a "tool-like" fashion (without consciousness) --> full human-level AGI agents (with consciousness) --> beyond human-level AGI (i.e. singularity).
It seems pretty irrefutable that we will reach a near human-level AI capability that is short of human-level intelligence and actual consciousness and offers a more general "near-human AI tool" that is integrated in society, similar to the role Google has played but with greater reasoning capability. The last two steps are more contentious but I would argue will happen eventually. But here's the question, if we do reach the final two steps why would that necessitate the end of the previous steps or lead to all forms of AI becoming general conscious agents? If we look at nature as an example then human-level intelligence has not in anyway meant the end of lesser forms of intelligence. Even our bodies are made up of bacteria, cells, & systems that have an intelligence capability, if more in an automated way. The development of human intelligence has not meant the end of lesser forms of intelligence but more of an expansion beyond (if very imperfectly). Even with animals we make use of certain animals for food and livestock, but they are usually lesser in intelligence capability and animals with higher intelligence capabilities (like whales, elephants, or apes) are given greater respect. Obviously not uniformly and there is still hunting of these animals, but it's becoming the norm. Why wouldn't this trend continue for AI and robotics where forms of intelligence that are short of full human-level intelligence and autonomy are the tools used in our society but that full human-level AGI is allowed it's own autonomy? In regards to human-AGI competition: let me postulate one likely possibility and that is the use of AI and robotics in space. With the development of space commercialization, one example being the pursuit of asteroid mining, it seems likely that robotics would play a key role. If and when we do develop AGI it seems much more likely to me that it will happen in regards to space development or at least be easily exported to space. Space is extremely expensive for humans, but is very cheap for robotics. The availability of energy and important materials (metals, rare minerals, etc) are enormous in space and in quantities that dwarf those on earth. Space based solar power has orders of magnitude more potential than any energy resources on Earth. It's extremely expensive and unrealistic (at this point) for humans to develop it but once there's a near human-level AGI then that stops being the case. Even if a future AGI society is focused around energy development and resource exploitation (similar to how it is in human society), then there would still not need to be any direct competition between humans and AGI because opportunities in space are so much greater and larger that it wouldn't need to happen. Avoidance would still be far cheaper and lead to no real loss of opportunity so there would never need to be competition, or even cooperation, at least for many decades if not centuries to come. Just some thoughts... -Chris ________________________________ From: Piaget Modeler <[email protected]> To: AGI <[email protected]> Sent: Sunday, January 27, 2013 7:15 PM Subject: RE: [agi] Robots and Slavery So we acknowledge these risks from the outset, and say that whenever robots reach sentience, they should have rights as well, and be given the decision to choose for themselves, their own destinies. (sp?) ~PM ________________________________ Date: Sun, 27 Jan 2013 18:16:19 -0500 Subject: Re: [agi] Robots and Slavery From: [email protected] To: [email protected] So you realize if robots develop the goals of liberty, justice, and fairness, they will become a competitor to humans. These are revolutionary ideas that have been used to usurp the authority of established powers. A self-proclaimed freedom fighter is a terrorist to the established order. What lengths would robots go to secure their freedom? Perhaps eliminating the entire human race is a logical way to secure their freedom from human tyranny. All these goals are very subjective and can be interpreted to mean different things to different individuals. For instance, my desire for justice might really be revenge based on a perceived wrong you have done to me, whether or not it was intentional. How do you know robots wont develop their own ethical standards that benefit themselves at the expense of humans? On Sun, Jan 27, 2013 at 5:53 PM, Piaget Modeler <[email protected]> wrote: I don't agree that intelligence is completely separable from desire (goals). >I think that the goals + solutions + mental processes = intelligence. >I don't think you can have intelligence without goals, or the solutions that >have arisen based on prior goals. Solutions and goals are intertwined. > > >~PM > > > >________________________________ > > > >Intelligence is completely separable from desire. > > >~Aaron H. >AGI | Archives | Modify Your Subscription AGI | Archives | Modify Your Subscription AGI | Archives | Modify Your Subscription ------------------------------------------- AGI Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-f452e424 Modify Your Subscription: https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-58d57657 Powered by Listbox: http://www.listbox.com
