RichardKennaway comments on AI box: AI has one shot at avoiding destruction - what might it say? - Less Wrong

18 Post author: ancientcampus 22 January 2013 08:22PM

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

Comments (354)

You are viewing a single comment's thread. Show more comments above.

Comment author: RichardKennaway 23 January 2013 12:34:50PM 8 points [-]

If the gatekeepers have such a high prior that the AI is hostile, why are we even letting it talk?

The point of the game is that there are people who think that boxing is a sufficient defence against unfriendliness, and to demonstrate that they are wrong in a way more convincing than mere verbal argument.

What are we expecting to learn from such a conversation?

In role, the gatekeeper expects to get useful information from a potentially hostile superintelligent being. Out of role, Eliezer hopes to demonstrate to the gatekeeper player that this cannot be done.