[ExI] Thousands of scientists pledge not to help build killer AI robots

[John Clark]

On Sat, Jul 21, 2018 at 9:00 PM, Stuart LaForge <avant at sollegro.com> wrote:

>
​>​
>
>
> *Killer AI robots are bad karma. Turing's Halting Problem pretty much
> makes it impossible to predict if or when a killer AI robot will ever stop
> killing.​ Set one loose after a terrorist network and it could go on a*

*busy-beaver killing spree. *


​But humans are not immune from Turing's Halting Problem anymore than
computers are. General Curtis Lemay went on a Busy Beaver style killing
spree in 1945, in the firebombing of of Tokyo he killed killed at least
125,000 people. It was probably the bloodiest 6 hours in human history and
he used nothing but dumb bombs, there was not a smart one in the bunch.

 Hell Killer AI robots would be an excellent
> ​
> candidate for the Great Filter.


I disagree. The puzzling thing is the apparent lack of intelligence in the
universe except on this planet
​,​
and robots have intelligence.

​>​
> *I suggest we always keep a human conscience on the trigger.*
>

​I don't see how that would help.

John K Clark


 ​
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20180722/9831a406/attachment.html>