[ExI] Eliezer Yudkowsky New Interview - 20 Feb 2023

[spike at rainier66.com]

From: extropy-chat <extropy-chat-bounces at lists.extropy.org> On Behalf Of Gadersd via extropy-chat
…

 

>>Why would you give a super intelligent AI the unchecked power to do potentially catastrophic things?

 

>…Because it’s profitable to give AI the authority to perform tasks traditionally done by humans. A super intelligence can potentially do quite a lot of harm with relatively little authority. A super intelligent hacker only needs to find a basic software bug to gain access to the internet and imagine what might happen next…

 

 

 

We could give AI jobs, such as whoever writes letters of condolence to people who have suffered tragedy on campus for instance.  There are entire departments dedicated to that sort of thing, but we have seen it can be easily, effectively and cheaply automated.  Software can create artificial empathy indistinguishable from the biological brand.  So… hire it.

 

But in all this, keep in mind that bad guys do not need AI to do bad acts.  Ordinary current Chatbots are now advanced enough to do bad acts.  For instance, China has access to the same internet we do, so they can get on ChatGPT and learn how evil the commies are who run their country.  OK bad example, that would be a good deed.  But the Chinese can also create chatbots and train them to teach capitalist youth how great communism is.  Conclusion: we need not wait for AI: ordinary software can already be used as a weapon.

 

Oh hey, now we need an arms race!  On your marks… get set…

 

spike





 

 

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20230228/b7d848eb/attachment.htm>