Less Wrong is a community blog devoted to refining the art of human rationality. Please visit our About page for more information.

TimoRiikonen comments on AGI Safety Solutions Map - Less Wrong

10 Post author: turchin 21 July 2015 02:41PM

You are viewing a comment permalink. View the original post to see all comments and the full post content.

Comments (7)

You are viewing a single comment's thread.

Comment author: TimoRiikonen 23 July 2015 12:47:16PM *  1 point [-]

Multilevel Safe AGI

I didn't quite understand this one. I am not sure if I understood this correctly.

Suggested updates if I did understand it:

1) There is one Nanny (although it may have multiple copies of itself, but basically one evolution path only)
2) AGIs must give Nanny access to verify that they are and remain peaceful
3) Separate which rules are on Nanny and which are valid to other AGIs...
Comment author: turchin 23 July 2015 01:04:26PM 0 points [-]

Multilevel Safe AGI is the way of integration of different solutions which will make them more stable.

Basically, what you are suggesting is under the topic " Use AI to create Safe AI". I may call it "Nanny Judge AI".