On Friday 07 March 2008 05:13:17 pm, Matt Mahoney wrote: > How does an agent know if another agent is Friendly or not, especially if the > other agent is more intelligent?
See Beyond AI, p331-2. What's needed is a form of open source and provable reliability guarantees. This would have to be worked out in great detail by the AIs themselves, but it would clearly be a very valuable thing for two AIs to be able to exchange trustability guarantees as part of a contract, so if we can't figure out how to, they probably will. As for people, it seems likely that a very valuable thing (for the individual and for the rest of society both) for each person to have a "Jeeves" AI which helps him navigate the complexities of AI society (parsing guarantees, for example) and also guarantees the human's behavior (and acts to enforce the guarantees if necessary). Josh ------------------------------------------- agi Archives: http://www.listbox.com/member/archive/303/=now RSS Feed: http://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: http://www.listbox.com/member/?member_id=8660244&id_secret=95818715-a78a9b Powered by Listbox: http://www.listbox.com
