[ExI] favor for a friend

[Adrian Tymes]

On Thu, Mar 25, 2021 at 10:38 AM spike jones via extropy-chat <
extropy-chat at lists.extropy.org> wrote:

> Adrian, never mind the images.  If we accidentally spawn an evil AI,
> images are the least of our worries.
>

On that front, I'm far more concerned about AI-enabled evil humans than an
evil AI.  There is ample evidence to show that humans can supply pure,
refined evil.

Imagine a servant AI that can do any well-enough-defined task almost
instantly, so long as it's only dealing with pure information that it has
access to.  Now imagine an evil - genocidal, perhaps - human who gets very
proficient about imagining and defining such tasks, including having the AI
research and find what information it can access.  This possibility isi
much closer than an AI that is evil in and of itself, no?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20210325/94b02a42/attachment.htm>