Wouldn't that be fairly easy to detect because the accounts would belong to an isolated, tightly-connected cluster?
Social graph metadata. The GGP said they would all be "talking to each other."
I think by 'detect' they mean 'detect by people who might fall for it' not 'detect by national security agencies.'
I don't see any way for the average social media user to detect whether they're engaging with a botfarm unless the social media company exposes their analytics, and they don't, and they aren't likely to start unless forced.
I actually was thinking the people who'd do the detecting would be the social media site operator (e.g. Twitter), who would then shut down the inauthentic activity.
You've just described half of Twitter's MAUs.
The TL;DR is that once it had enough compromised machines to run social media botnets, it was all over. It could use those to confound efforts to coordinate and compare data, to misdirect huge numbers of people and cause all kinds of chaos, and to smear opponents before they could get their message out (fakes or actual stolen information—it hardly mattered, all it needed to do was neutralize certain people for a few days). The story contrived to have a secret project that was able to try to resist it after that (spoiler: didn't help) but otherwise the social media botnets were enough for it to buy several days in which no-one was able to effectively work against it.