Re: Truth machines

Michael Lorrey (mike@lorrey.com)
Thu, 17 Dec 1998 12:36:01 -0500

Eliezer S. Yudkowsky wrote:

> JKC has noted that evolution has put a tremendous amount of effort into
> detecting and concealing lies, so a perfect truth machine is unlikely.
>
> I say exactly the opposite: Evolution has put so much effort into lies
> that there's probably a module of the brain devoted to lying (anyone
> know if someone's looked for it?), and thus it might be very easy to
> detect activity with an fMRI. Evolution baffles verbal and kinesic
> perceptions, but would have absolutely no reason to defend against
> neuroimaging. The inventor might not even need much cognitive science;
> a neural net might be very easily trainable to decode "lying" brain activity.

Yes, there was an interesting bit on tv recently on the pantomimes that people have when they are lying. For example, apparently the memory center of the brain is in one hemisphere, while the center of creative thinking is in the other. You can tell which one the person is accessing by watching which way they look, to the right or left, when they talk. If they are lying, then they are accessing the creative center to make the story up. This trick can be beaten by creating your story before hand, and memorizing it, so you can remember it when being questioned, rather than making it up on the fly, but there are a number of other signals as well. Making a perfect thruth machine is merely a matter of having sufficient sensors. You can fool some of the sensors some of the time, but not all of them all of the time.....

Mike Lorrey