On Tue, 25 Apr 2023 10:53:06 -0000 (UTC)
Spiros Bousbouras <spibou@gmail.com> wrote:
> On 25 Apr 2023 01:30:07 -0300
> Mike Spencer <mds@bogus.nodomain.nowhere> wrote:
> >
> > Spiros Bousbouras <spibou@gmail.com> writes:
> > > This discussion has made me wonder though whether AI will be used
> > > for usenet trolling. People certainly have triggers on usenet (for
> > > comp* related , a typical example would be <programming language A
> > > vs programming language B>) and an AI which is sufficiently well
> > > tuned and trained , may be able to keep such discussions (flamewars)
> > > going indefinitely.
> >
> > Why limit it to Usenet?

Because I consider unmoderated usenet exceptionally susceptible to this
threat. For example one could use a few AIs which endlessly crosspost on
comp.lang.c and comp.lang.c++ discussing which is the best language. One
doesn't have an obvious way to quickly filter this kind of thing. The posts
would be polite , erudite and generally make reasonable arguments and many
humans would bite and participate. But such discussions could be made to
dominate the groups until eventually many humans become fed up and stop
reading. I don't think it would take large computational resources either ,
probably within the reach of an average individual now or shortly in the
future. I'm actually surprised why even human trolls don't seem to have done
this much.

--
If I had a share of AOL for every time someone told me that the web would die
because AOL was so easy and the web was full of garbage, I'd have a lot of AOL
shares.
And they wouldn't be worth much.
https://boingboing.net/2020/01/27/nascent-boulangism.html

> Because I consider unmoderated usenet exceptionally susceptible to this
> threat. For example one could use a few AIs which endlessly crosspost on
> comp.lang.c and comp.lang.c++ discussing which is the best language. One
> doesn't have an obvious way to quickly filter this kind of thing. The posts
> would be polite , erudite and generally make reasonable arguments and many
> humans would bite and participate. But such discussions could be made to
> dominate the groups until eventually many humans become fed up and stop
> reading. I don't think it would take large computational resources either ,
> probably within the reach of an average individual now or shortly in the
> future. I'm actually surprised why even human trolls don't seem to have done
> this much.

Agreed, and there are precedents for such crapfloods, like the stupid
Meow Wars of 1996-1998 (https://en.wikipedia.org/wiki/Meow_Wars).

Whoever invents such a system would probably see the likes of Twitter
or Reddit a more lucrative/enjoyable target than Usenet though. Hoping
the relative low volume of Usenet makes it a not very interesting
target.

Spiros Bousbouras <spibou@gmail.com> wrote:
>Because I consider unmoderated usenet exceptionally susceptible to this
>threat. For example one could use a few AIs which endlessly crosspost on
>comp.lang.c and comp.lang.c++ discussing which is the best language. One
>doesn't have an obvious way to quickly filter this kind of thing. The posts
>would be polite , erudite and generally make reasonable arguments and many
>humans would bite and participate. But such discussions could be made to
>dominate the groups until eventually many humans become fed up and stop
>reading. I don't think it would take large computational resources either ,
>probably within the reach of an average individual now or shortly in the
>future. I'm actually surprised why even human trolls don't seem to have done
>this much.

This is not a new threat. AI is not needed (whatever AI is), nor is any
machine learning. You can just create messages from a list of different
arguments and sentences, just re-arranging them all the time. It would be
detectably no different than what some of the regulars in those groups post
anyway; the only difference would be in volume.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."