Social graph of mysterious twitterbots


#1

[Permalink]


#2

Whatever measures are taken to fight this will just further spur evolution of the bots’ ability to pass as human. #storyidea


#3

Turing Test as selective pressure?


#4

Model of a neural network?


#5

But why? What do bots get out of passing for human? We already have humans…


#6

Cory or Rob or anyone else with the keys: sorry for the hijack but has the format formerly accessible at boingboing.net/page/1 been retired, or is it just down temporarily? I can’t access it currently, and find it much preferable to the default view. Thank you.

EDIT: it’s back, thank you to the server Gods and/or Cory/Rob/anyone else who fixed it.


#7

And how well has that been working out for us? Huh?

That’s what I thought.


#8

You are guessing that you already know my answer to your vague question?

It works fine. I just prefer the company of people who aren’t confused, and it seems to me that those who aren’t confused are more likely to take things for what they are, rather than pretend that they are something else. Bots are fine, and people are fine. But trying to make one out to be a stereotyped representation of the other? Not so much.


#9

So in a delightfully meta twist on the subject, some of the bots have begun tweeting to your previous post about them Cory. See the second link in: https://twitter.com/BoingBoingHuffR/status/574356165506777088


#10

Ooooh, that’s good. I’ll need to remember that one.


#11

Wondering if there aren’t a bazillion similarly automated accounts on Facebook. I’ve noticed among my conservative friends a recent upswing of Fox News-like posts, which are always accompanied by hundreds of thousands of likes. It’s not that those kinds of numbers are impossible, but… I wonder.


#12

I think alot of these things are abandoned spammer bots just following their own tales now. Thousands of fake accounts run from compromised computers spreading themselves with no target anymore.

Not sure if that’s what these are but I know there are bots out there like what I describe.


#13

I think the number of actual Facebook users are somewhere between half and one third what Facebook claims. So many fake SEO accounts, bots and just normal people who have multiple accounts. I get made in infosec classes when professors just blindly quote the 1 billion Facebook users number.


#14

…many of whom could easily pass for bots.


#15

I can’t say why they have lots of fake followers…

Wouldn’t taking a look at their content give us a clue to their purpose? For example, if they’re incessantly tweeting, “drink your Ovaltine…”


#16

The bots are out to steal jobs and government benefits from honest, tax-paying humans.


#17

It is very fascinating - and has clear commercial value. If one can manipulate the news through apparent reports then it can have an impact on prices, either commodities or equities. The Louisiana chemical plant story, for instance, would increase prices of whatever the plant was producing and depress feedstock prices.

Foreknowledge of such an event would allow market participant(s) to benefit from these fluctuations, in addition to shorting the stock of whoever owned the apparently targeted plant.

One could look at how the hacking of social media accounts belonging to major media houses or governments has had an impact. Such as the AP hack in April 2013, which caused the Dow to drop 100 points for a couple of minutes. This is the larger end of the spectrum - and was quickly scotched - but smaller, more focused “false news” would be slower to be denied and benefits for the news manipulators more pronounced.


#18

Or, maybe it demonstrates that commerce is automatic enough to not be worth people’s time or effort! Why let an algorithm change prices for people, when you could instead have the algorithms earn and spend the money, buy and sell the products - all in the box - while humans go do something more interesting? This is where the real possibility lies. ONLY computers using money and commerce.


#19

My software tells me that I am likely to find your ideas intriguing; I have allowed a script to subscribe to your newsletter.


#20

I am curious too, but take them as “government”, and that of “five eyes” variety. (Or ‘12 eyes’ or ‘19 eyes’ or whatever, with an emphasis right now, on GCHQ UK’s NSA).

So, from that consideration… I think they are going to operate by trial and error, improving their systems … until… they get it right.

In attempting to do so, maybe they will find some practical limitations, and so stop on certain niche uses. Or, maybe something truly sinister might result.

Though… it can well be said the premise its’ self is already sinister in conception. Not at all unlike the 50 Cent army of China, where mass Chinese go out there and propagandize, censor, and report under the auspices of that golden nickel.

But, is it effective? After all, do not groups naturally already behave in this way, and is not true art inherently manipulative by the very fact of it being so profoundly good at communication?

There is, though, a distinct difference between opinion which is properly subjected to at least an attempt at running through a system of proper weighting. And a system which attempts to derail that honest process of accurate measurement. It is as dumping into the economy of trust counterfeit material designed specifically to destroy flourishing economies of opinion. And the counterfeit nature of it can be seen exactly as it is: worthless junk that is inherently dangerous by its’ destructive capacity.