Clearly if they exist they must have some marginal use. It's like saying "who falls for nigerian prince scams?" or "who is influenced by ads?". Clearly some people are, otherwise they wouldn't be quite as popular. I'm sure some real people follow porn bots in order to... see porn.
As for the influence I think in isolation it's probably negligible, but if people get flooded by propaganda on social networks it may be quite efficient. "Everybody says it, it can't be completely wrong". The echo-chamber nature of many social networks probably plays into that too. Many people would probably be more likely to follow an obvious bot spamming propaganda that aligns with their beliefs than a real person holding a different opinion. For instance an obviously biased propaganda outlet like Prager U has 2 million followers on Youtube, I doubt all of those are bots.
Searching for contradictory viewpoints and actually managing to listen to the argument with an open mind takes effort. That's the big problem with most social media currently, they amplify this bias to the extreme. You can hold a fringe belief like "earth is flat" and still manage to find thousand of people agreeing with you, spending hours every day consuming flat earth content and avoiding any dissenting voice.
The article says accounts like this can be purchased cheaply from resellers.
"How does that influence anybody?"
I think the most effective use is in replies to other tweets. All the replies to a tweet are shown if you look at a tweet. You might look at the profile of some user who makes a tweet, but perhaps won't take such a deep look at their profile to see the earlier porn tweets.
High number of tweets, high number of followers, long time on the site would indicate 'actual person' rather than bot, or so you'd think without scrolling down far enough to see it filled with spam posts.
In my view bad actors don't have to just misinform, but make it more difficult to locate. Doing that seems shockingly easy, as much of our ability to locate data online is limited to brittle keywords.
Not to deviate, but in case anyone is curious: One of my attempts at mitigating this problem is via content hashes. Building a UX where the content hashes are often the point of reference and indexed by search engines might allow us to locate data and related data based on content hashes, rather than easily flooded semantic keys. Metadata pertaining to the hashes would also have to be provided for them to be useful, eg "related:04136e24". Lots of to discuss, but I'm trying to limit my off-topic, despite it being related to [dis|mis]information.
edit: I don't understand the downvote - can someone explain? Not sure exactly if something I said is factually incorrect? To be clear I wasn't saying Google was the thing running the bots, but that Google / search engines might be the point.
Not much different from the link farms of the early days of SEO which were never meant for human eyes.
Less about the credibility of the account, and more about creating a lot of buzz or, well, counter-buzz around another thread.
So there are bots that have been posting "porn" for a bunch of years. And now they post political messages.
How does that influence anybody? Who subscribes to porn bots? I would have thought only other porn bots follow porn bots?
Can somebody link to an example of such a bot?