It’s happening not just in USA. In Slovakia there was recently a research published about fake accounts commenting on social networks and comment sections of major newspapers. It’s quite scary how effective the guerrilla information war can be. This is how it works, in a nutshell:
You hire a few people (you really don’t need many of them) working for minimal wage, speaking the language of the target country. Their task is to create dozen accounts on various websites. If names and photos are required, they just make fake names and steal someone’s photo from google images. (If someone finds that out and bans the account, no problem, they create a new one).
Separately from these people, you have a few smart guys who decide what memes they want to spread. They choose the message, compose a few dozen text samples saying the same thing using different words. Then they e-mail the text samples to the minimum-wage people, whose task is essentially to keep copy-pasting these messages on many places with minor modifications, for 8 hours a day.
Imagine a group with e.g. one manager, two smart guys designing the memes, and five stupid guys who copy them. How many comments they can generate each day? How much can they influence a perception of an average reader?
That form of spamming hundreds of websites by minor modifications of messages optimized for an average reader obviously wouldn’t survive on LW. And optimizing messages specifically for LW wouldn’t be cost-effective. The operation needs to scale well; there is a whole internet to take over, and even Kremlin’s propaganda budget has its limits.
But I could imagine a multi-tiered approach. Something like categorizing the websites into a few groups, and crafting messages for each of these groups separately. The categorization itself could actually be quite simple: for each category, you choose a few “prototypical” websites, and then let a neural network assign the right category for everything else. Then you do the same thing—have a smart guy writing a message optimized to impact people in given category, and a stupid guy to spam all websites assigned into given category with small modifications of the message—for each category separately.
With enough categories, I could imagine that one of them could be compatible enough with LW. Probably “rationalists” is too narrow, but something like “educated people” or maybe even “STEM” could get among the first dozen categories.
It’s happening not just in USA. In Slovakia there was recently a research published about fake accounts commenting on social networks and comment sections of major newspapers. It’s quite scary how effective the guerrilla information war can be. This is how it works, in a nutshell:
You hire a few people (you really don’t need many of them) working for minimal wage, speaking the language of the target country. Their task is to create dozen accounts on various websites. If names and photos are required, they just make fake names and steal someone’s photo from google images. (If someone finds that out and bans the account, no problem, they create a new one).
Separately from these people, you have a few smart guys who decide what memes they want to spread. They choose the message, compose a few dozen text samples saying the same thing using different words. Then they e-mail the text samples to the minimum-wage people, whose task is essentially to keep copy-pasting these messages on many places with minor modifications, for 8 hours a day.
Imagine a group with e.g. one manager, two smart guys designing the memes, and five stupid guys who copy them. How many comments they can generate each day? How much can they influence a perception of an average reader?
Thanks for the information! May I get a link to that research paper? My google-fu is apparently weak.
Sorry, I have a problem finding it now too. :(
Here are a couple of useful links—from the New York Times and the Guardian.
Information wars are quite real. How much actual impact do they have is a different question.
A fun possibility to consider: do we have Russian mole here on LW? X-D
That form of spamming hundreds of websites by minor modifications of messages optimized for an average reader obviously wouldn’t survive on LW. And optimizing messages specifically for LW wouldn’t be cost-effective. The operation needs to scale well; there is a whole internet to take over, and even Kremlin’s propaganda budget has its limits.
But I could imagine a multi-tiered approach. Something like categorizing the websites into a few groups, and crafting messages for each of these groups separately. The categorization itself could actually be quite simple: for each category, you choose a few “prototypical” websites, and then let a neural network assign the right category for everything else. Then you do the same thing—have a smart guy writing a message optimized to impact people in given category, and a stupid guy to spam all websites assigned into given category with small modifications of the message—for each category separately.
With enough categories, I could imagine that one of them could be compatible enough with LW. Probably “rationalists” is too narrow, but something like “educated people” or maybe even “STEM” could get among the first dozen categories.