|
|
|
|
|
|
by dogma1138
2682 days ago
|
|
|
I’m not entirely sure that that bad actor would get any more scalablity form it than from a Mechanical Turk farm, at least as far as impact goes. It seem that as far as information warfare goes “less is more” works quite well and they rely on targeted people to spread the news for them. When you want to drive an agenda you don’t need unique 100,000 comments you need a good copy pasta. Overall I’m sick of this dramatization of the AI catastrophe until there will be a proven path with agency for it to actually operate in the real world. A chat bot isn’t a threat to anyone even if it turns homicidle. |
|
|
Regarding an agenda, sure, good pasta is fine and all, and regular ol people are fine, but it is not cost effective. This is a million times cheaper, which means you can use it everywhere, not just the obvious places, you can be everywhere, and you can do more than just push a couple big items, you could push tens of thousands of them, micro targeted all the way down to the individual. Don’t dismiss it so easy—the potential scale is far, far larger than anything existing to date.
And I would note that the reason 100,000 comments aren’t effective now is precisely because they are too formulaic, too obviously fake when used on such a large scale. This has the potential to create real, live, seemingly active and believable online communities of millions of people, all at fractions and fractions and fractions of a penny compared to current methods. People read news, then comments (or reviews or whatever), because they use them to determine the validity of the content they just read; if it’s no longer possible to tell from the comments what’s a scam and what isn’t... well, you could do a lot of things with that.