Join 3,496 readers in helping fund MetaFilter (Hide)


OK, time to own up.
February 23, 2011 9:49 AM   Subscribe

Online astroturfing is more advanced and more automated than we’d imagined.

"Astroturfers can then be assigned 'pre-aged' accounts to create a back story, suggesting that they’ve been busy linking and re-tweeting for months... "

"With some clever use of social media, astroturfers can 'make it appear as if a persona was actually at a conference and introduce himself/herself to key individuals… There are a variety of social media tricks we can use to add a level of realness to all fictitious personas.'"
posted by randomination (114 comments total) 41 users marked this as a favorite

 
Oops, some of my punctuation went awry there. Sorry, corporate masters.
posted by randomination at 9:51 AM on February 23, 2011


Previously.
posted by kagredon at 9:52 AM on February 23, 2011


Excelent story . I agr'ee ! Good jorb !
posted by Plutor at 9:54 AM on February 23, 2011 [10 favorites]


Hey guys, ever wonder what's going on with Digg these days?

...oh...
posted by 2bucksplus at 9:58 AM on February 23, 2011 [28 favorites]


Looks like HBGary has had a lot more work than we imagined.
posted by hippybear at 9:58 AM on February 23, 2011


Shades of Orwell's 1984.
posted by ZenMasterThis at 9:59 AM on February 23, 2011


Shit, I forgot to log out all the personas... sorry for all the one liners
posted by infini at 10:03 AM on February 23, 2011


I hope this will stop people bitching and moaning about how Anonymous has never done anything useful or politically-meaningful.
posted by nasreddin at 10:04 AM on February 23, 2011 [10 favorites]


I always wondered who those folks supporting Apples increasingly shady business practices were...
posted by zeoslap at 10:05 AM on February 23, 2011 [12 favorites]


That's interesting you say that about Apple. Say, did you see this article about the exciting new upcoming iPad?
posted by Cool Papa Bell at 10:08 AM on February 23, 2011 [5 favorites]


That's interesting you say that about the iPad. Say, did you see this article about the exciting new upcoming Motorola tablet?
posted by Cool Papa Bell at 10:08 AM on February 23, 2011 [5 favorites]


That's interesting you say that about the tablet. Say, did you see this article about the exciting new drug from Pfizer?
posted by Cool Papa Bell at 10:09 AM on February 23, 2011 [18 favorites]


I guess that explains dhoyt et. all.
posted by Freen at 10:10 AM on February 23, 2011


NOW CUT THAT OUT.
posted by oneswellfoop at 10:10 AM on February 23, 2011


NOW CUT THAT UP.
posted by oneswellfoop at 10:10 AM on February 23, 2011


NOW CUT THAT WITH THESE AMAZING GINSU KNIVES.
posted by oneswellfoop at 10:10 AM on February 23, 2011 [23 favorites]


You are not nearly as clever as you think you are.
posted by nasreddin at 10:12 AM on February 23, 2011 [1 favorite]


Is this why any time Israel does something awful, the Fifty Cent brigade comes in to defend them? I hope I'm wrong.
posted by dunkadunc at 10:12 AM on February 23, 2011 [1 favorite]


DERAIL HERE, DERAIL NOW
posted by kagredon at 10:17 AM on February 23, 2011 [2 favorites]


MetaFilter's $5 entry fee is looking smarter and smarter. Financial barriers are the only methods I've seen work for large web communities when it comes to keeping out the astroturfers.
posted by Kattullus at 10:18 AM on February 23, 2011 [41 favorites]


{THREAD DESCENDS INTO BIG NASTY ARGUMENT WITH TROLLBOTS}
posted by dunkadunc at 10:18 AM on February 23, 2011 [1 favorite]


*sigh*

*places thread back onto rails*

Now where were we
posted by randomination at 10:18 AM on February 23, 2011 [3 favorites]


Now where were we

It's interesting that ask where you are. Have you checked out the latest deals on travel packages?

OK, OK, I'll stop...
posted by Cool Papa Bell at 10:20 AM on February 23, 2011


That actually sounds kind of frightening.
posted by neuromodulator at 10:22 AM on February 23, 2011 [2 favorites]


I say we purge the sockpuppets.
posted by clavdivs at 10:22 AM on February 23, 2011 [1 favorite]


The nominal fee to register an account solution has worked pretty well here, but someone sufficiently financially motivated could just as easily pony up $500 and get a hundred-member strong sockpuppet army.

However, I think this is a solvable problem. If people are creating and selling "persona management" software then there are going to be similarities between those accounts, particularly with the fake "pre-aging." Not all timestamps are going to be spoofable in an undetectable way. So, it should be possible to write astroturf-detecting software which could identify the sockpuppets. It's even easier if you're already charging a fee to register an account, you just allow a maximum number of transactions per paypal account or whatever.
posted by signalnine at 10:22 AM on February 23, 2011 [2 favorites]


Randomination is obviously in the pockets of the railroad barons.

In all seriousness, while this is a problem, I'm not sure what the lay writer/reader/commenter on the web can do to prevent or stop this. Without dedicated, technologically sophisticated investigation, how do you distinguish an astroturfer from a true believer (or a troll?)
posted by kagredon at 10:23 AM on February 23, 2011


MetaFilter's $5 entry fee is looking smarter and smarter. Financial barriers are the only methods I've seen work for large web communities when it comes to keeping out the astroturfers.

Even this won't stop a company determined to astroturf. $5 an account starts to look more like an investment than a hurdle to jump over.
posted by JHarris at 10:24 AM on February 23, 2011 [4 favorites]


Impose a syntax.
posted by clavdivs at 10:24 AM on February 23, 2011 [12 favorites]


We could just write software to filter out anyone who expresses an opinion.

Not that I'm advocating that. It's just an idea.
posted by AugieAugustus at 10:24 AM on February 23, 2011 [3 favorites]


From the article: "Comments are closed, but trackbacks and pingbacks are open."

That's the only way to keep out the astroturfers. But the $5 keeps 'em to a minimum.
posted by blendor at 10:26 AM on February 23, 2011


MeFi also has good moderation - checking IPs etc.. The other solution is FaceBook where you can verify who the poster is (non-anonymous).

Those are the two ways forward: moderation and/or non-anonymous. Entrance fees? Maybe but as others point out $500 for 100 accounts is very little money for a PR firm backed by oil clients.
posted by stbalbach at 10:28 AM on February 23, 2011


Is this why any time Israel does something awful, the Fifty Cent brigade comes in to defend them? I hope I'm wrong.

You're not.
posted by ymgve at 10:28 AM on February 23, 2011 [14 favorites]


Software like this has the potential to destroy the internet as a forum for constructive debate. It makes a mockery of online democracy. Comment threads on issues with major commercial implications are already being wrecked by what look like armies of organised trolls – as you can often see on the Guardian’s sites.

No it doesn't, because internet users can and do discount the input from anonymous or otherwise unverifiable contributors. 'I heard it from someone on the internet' equates to 'unreliable source' for most people. Aggravating though it is, the phenomenon is fundamentally no different from a politician or lobbyist packing a public meeting to create the appearance of broad public support for a particular viewpoint.

There is no simple technical solution. Every few years someone proposes introducing an 'internet driving license' or placing statutory limits on online anonymity. However, any security system can be subverted and one intended for the use of the general public will have to be so open as to be easily subvertible. The existence of such a system would be a blessing to trolls because it creates a false sense of security, while making it easier to identify and abuse the privacy and security credentials of less tech-savvy individuals. Anonymity is a fundamental component of privacy; for those who approve of 4chan's civil disobedience under the banner of 'anonymous,' it's worth bearing in mind that such activities would be the first to go. A better approach (in my view) is the development of tools to correlate and track unusual patterns of posting, eg by observing that 5 supposedly distinct individuals have a remarkably consistent habit of saying the exact same things on 25 different websites day after day.

Of course both the persona and persona detection tools are capable of being used for good or ill. The idea that the Air Force might use such things doesn't inherently bother me, because I can think of some legitimate intelligence-gathering or even propaganda purposes - on the other hand, I can think of many illegitimate ones, so I would like to know more about why the USAF wants such tools. I'm also aware of private political groups employing techniques like this, with or without the help of special technology. This bothers me, but on the other hand the semi-automated nature of the astroturfing is what helped me to spot the behavior in the first place (in relation to a particular issue that I'm interested in).
posted by anigbrowl at 10:31 AM on February 23, 2011 [5 favorites]


stbalbach: "Maybe but as others point out $500 for 100 accounts is very little money for a PR firm backed by oil clients."

That's nothing. Big Turpentine paid like $10 for my account years ago...
posted by l33tpolicywonk at 10:32 AM on February 23, 2011 [1 favorite]


stbalbach: The other solution is FaceBook where you can verify who the poster is (non-anonymous).

You can create fake Facebook accounts fairly easily. In Iceland there have been several fake Facebook accounts have been brought to light, but it would be impossible to use the methods used in Iceland in larger societies (i.e. looking people up in the genealogical database of all Icelanders).
posted by Kattullus at 10:39 AM on February 23, 2011 [2 favorites]


A better approach (in my view) is the development of tools to correlate and track unusual patterns of posting, eg by observing that 5 supposedly distinct individuals have a remarkably consistent habit of saying the exact same things on 25 different websites day after day.



http://www.metafilter.com/100679/Rise-of-the-Sockpuppets
posted by infini at 10:39 AM on February 23, 2011


The other solution is FaceBook where you can verify who the poster is (non-anonymous).

Well, yes, but as we've all learned from the HBGary thing, there are people working to create social networking tools which allow you to create wholesale Facebook sockpuppets, too.

I can't find the article at the moment, but I read something very recently which laid out their strategy, even going so far as to talk about how they were going to create personae which people would believe. It involved choosing medium to large cities for the "person" to be from, creating a backstory which would have that person leaving shortly after school (military service was specifically suggested), and then working backwards from there to establish links with people who could have brushed up against that "person" in the intervening years, and then finally using the developed profile to start linking to individuals in the target social circle which was desired to be infiltrated.
posted by hippybear at 10:42 AM on February 23, 2011


I, for one, welcome our new automated overlords.
posted by fifteen schnitzengruben is my limit at 10:44 AM on February 23, 2011 [2 favorites]


I honestly think these tactics are ineffective as marketing, much as most internet marketing is driven by hype rather than results. However, they have the deleterious effect of making online communities viscerally unpleasant. People are remarkably willing to ruin things for the infinitesimal possibility that someone is going to fall for their marketing crap.
posted by sonic meat machine at 10:47 AM on February 23, 2011 [4 favorites]


The irony of having to use transparent tender bidding processes... poor dim.
posted by infini at 10:48 AM on February 23, 2011 [1 favorite]


I'm waiting for the AskMe that needs recipes using only Kraft brand food or some such (for a bet? a scavenger hunt? a theme party?)
posted by infini at 10:50 AM on February 23, 2011 [2 favorites]


If you've a better idea on how to get Milli Vanilli into the #1 slot, I'd like to hear it.
posted by RobotVoodooPower at 10:52 AM on February 23, 2011


Back in 1995/6, I was a cameraman for a local cable news program. One of my biggest assignments was covering the Al Sharpton defamation law suit brought by Steve Pagones in relation to Tawana Brawley. Every day, Sharpton's team trucked in the same 50 protesters from out of town on hired buses. EVERY. DAY. In other words, these people were EMPLOYED to stand outside and protest. My point being, astroturfing is simply using new tools to do the same old thing that entrenched interests and oppositional parties have done for ever and ever.
posted by spicynuts at 10:54 AM on February 23, 2011 [2 favorites]


sonic meat machine: "I honestly think these tactics are ineffective as marketing, much as most internet marketing is driven by hype rather than results. However, they have the deleterious effect of making online communities viscerally unpleasant."

Exactly. And if you want to shut down any productive discussion of a topic, just drop in with some robotrolls to spread some disinformation and say some Obviously Wrong Stuff that people will spend all their energy refuting.
posted by dunkadunc at 10:56 AM on February 23, 2011 [2 favorites]


You'll notice that 4chan /b/ requires answering a captcha for every single post. Anonymous started where again?

If that seems excessive, you could only demand captchas randomly, except that ip blocks that routinely fail would get asked for captchas all the time.

You could obviously circumvent simplistic capatchas by hiring chinese people to solve them, but that'll cost more than $5 and capatchas can evolve too.

Examples : reCapatcha could use the host country's language, thus protecting non-english speakers from English speaking capatcha solvers in China and India. Audio and image identification capatchas are more culturally sensitive too. If your looking for astroturfing ip blocks, even silent flash applets yield some information.

You might even try building LSAT style logic puzzles with the final phraseology constructed by a Markov generator and typeset into a capatcha image using reCapatcha's word database.    Evilest. Capatcha. ever!
posted by jeffburdges at 10:57 AM on February 23, 2011 [5 favorites]


Without dedicated, technologically sophisticated investigation, how do you distinguish an astroturfer from a true believer (or a troll?)

The more generic a person's responses, the higher the probability of astroturfing. Short generic comments are the hardest to detect. For example, if there's an anti-Santa campaign, comments like 'I hate that Santa guy!' and ''We the people' reject Santa!' are so simple as to be individually indistinguishable. But if someone posts a 150 word rant about Santa's oppression of the polar elves or Santa's secret communist agenda, that's linguistically distinctive. You can look for unusual combinations of words (known as statistically improbable phrases) and then use these as input to a search engine, which may show you that the same rant has been pasted word-for-word on 50 different forums. This is sort of how anti-plagiarism software works.

On the other hand, if someone is engaged in an argument on a site like MeFi and there's a lot of back-and-forth with other posters, that kind of interaction is time-consuming and doesn't easily translate to other contexts. You can't lift a long hairy argument from MeFi and just dump it into a discussion on FetaMilter.com (a cheesy website populated by fans of derailleur gears). It won't make sense over there due to the lack of conversational context, so it will be ignored or even rejected as copypasta. Thus, the astroturfer has little economic incentive to get involved in any but the most superficial discussion. Although a person could spend all day propagandizing one site (something I get accused of from time to time), putting a value on that activity is quite difficult, just as it's difficult to put a value on any single showing of an advertisement. It's easier for both the buyer and the astroturfer to focus on quantity over quality, because quantity can be easily measured and (to some extent) scored for effectiveness. Trying to measure the persuasiveness and influential power of one-off postings is almost impossible - like literary criticism, there's no clearly correct index of quality and so it has correspondingly lower utility for propaganda purposes except in highly specialized contexts.

I don't know of a simple software utility for easily tracking this sort of thing; my current approach (which is experimental and unstable) involves some screen-scraping and search engine API queries. I haven't spent a whole lot of time on it because I don't want to spend 3 months rebuilding development skills in order to reinvent the wheel and badly guesstimate the troll/astroturf/stoopid ratios in online news story comments. But I do think there's a lot of utility for software to track the spread of ideological memes in social media - suggestions would be welcome.
posted by anigbrowl at 11:00 AM on February 23, 2011 [12 favorites]


We have a different filter here. Thousands of members, all of whom are different (sockpuppets aside). We don't get to post until we have been here a little while. Obvious astroturfing is killed right quick and the permaban is in effect. The stuff that might slip by, well, it takes a long time and a lot of effort to cultivate a personality here that isn't going to be viewed with suspicion. Someone like DU makes a semi-spammy post, we might say "Pepsi Blue, delete" but nobody's going to go apeshit and call the Mefi Detective Squad down. Someone with a user number less than a few months old, little to no posting history and a very thin record of comments? Banhammer time.

This entirely does not work in most blogs where commenting is essentially a long list of people retweeting the article headline. That kind of shit isn't ever going to be safe from astroturfing. Who has the energy to moderate that?
posted by caution live frogs at 11:08 AM on February 23, 2011 [2 favorites]


I guess what I meant to add there (and didn't) is that MeFi is relatively safe from this simply because the structure of the site doesn't make the investment worth the outcome.

Man sometimes I just ought to post the tl;dr version up front... but if it wasn't overly wordy, how would you all know it was really me?
posted by caution live frogs at 11:09 AM on February 23, 2011 [1 favorite]


Just after reading this discussion, I read this article on Libya (warning, embedded videos are not for the squeamish).

There are several posters massively trolling and ranting about the President of the US. Perhaps they're just insensitive and like to stir up GRAR, but the comments are so unhinged in the context of the news article they're supposedly about. In this case, they're starting with an article about children and protestors being shot in Libya, and then making this about 'US Politician (or Political Party) is a fraud / clown / etc. I could easily see this being a bot-driven or cut and paste attack with minimum human intervention, as the comments have nothing to do with the original post.

I see this as matching the astroturfing approach mentioned by the OP:

1) take popular article
2) fill commentary with (political) position unrelated to article
3) use multiple personalities to have a 'conversation' about this position
posted by zippy at 11:09 AM on February 23, 2011 [2 favorites]


several posters on the linked discussion massively trolling
posted by zippy at 11:10 AM on February 23, 2011


> Just after reading this discussion, I read this article on Libya (warning, embedded videos are not for the squeamish).

Holy shit. I'm not squeamish but there are some seriously grisly pictures there.
posted by Burhanistan at 11:22 AM on February 23, 2011


Wow zippy,
That thread was something else. Is someone trying to set a new record for complete flamewar? Are Olympic qualifiers on right now?
posted by TheWhiteSkull at 11:38 AM on February 23, 2011


'internet driving license'

I, for one, would fucking love this. I'll miss you all, but whatever. I enjoyed the internet a lot back in '91 as well.

Test 1: build a TCP/IP stack on a microcontroller. Extra points of you implement it in VHDL and a FPGA.

Test 2: There is no test 2. You just built an internets.
posted by GuyZero at 11:41 AM on February 23, 2011 [5 favorites]


I guess what I meant to add there (and didn't) is that MeFi is relatively safe from this simply because the structure of the site doesn't make the investment worth the outcome.

then, there are the patterns, one week old user with 3 comments and a container
posted by infini at 11:46 AM on February 23, 2011


Would it be possible to create speech honeypots, like websites and twitter accounts designed to draw out autocommenters?
posted by drezdn at 11:53 AM on February 23, 2011 [10 favorites]


I enjoyed the internet a lot back in '91 as well.

Me too. I still have revenge fantasies involving Cantor & Siegel, although I'd like to state publicly that these involve nothing more harmful than the comical misuse of cream pies. But for all the problems, I don't think mandatory user tracking is the way out of the tunnel.
posted by anigbrowl at 11:57 AM on February 23, 2011


With any form of advertising, sooner or later the majority of people learn to recognize when they're being marketed to. It happened with product placement in movies (who doesn't laugh when the main character purposefully holds a coke with the label positioned just so?). So who can't we trust now? People with particularly strong opinions. People who advocate for big brands. The person who says, "yeah, Coke is good but Joe's Iced Tea Soda is better" will be more trusted because Joe probably won't be able to afford astroturfing (or Joe is a Coke subsidiary). And people who have online presence going back before the astroturfing era (making me glad I've been lax with my identity since the days of Compuserve, I guess). But how long until a Coke rep comes along and offers me $50k to promote their vile chemical swill online?
posted by TochterAusElysium at 12:09 PM on February 23, 2011


I like that idea, drezdn.

Or your own bots that find botted comment threads so they can comment with an explanation of the astroturfing at hand.

BOTWARS!
posted by defenestration at 12:10 PM on February 23, 2011


Oh, Anonymous. I should've known you played both sides.
posted by majonesing at 12:19 PM on February 23, 2011


China has been doing this for a long time, with the '五毛党 (wǔmáo dǎng) or '50 cent party members' who supposedly get paid 5 mao to comment on the internet for the government.

I bet anything the telecom companies have been doing this to badmouth and sew doubts about net neutrality.
posted by delmoi at 12:24 PM on February 23, 2011


I trace this back to 1994, but I am sure somebody somewhere was doing it soon after the internet was first connected.

For a period of several months in the first half of 1994, the Internet user under the pseudonym of "Serdar Argic" (with the address sera@zuma.uucp) posted messages in any Usenet newsgroup thread involving the country of Turkey, arguing that the Armenian Genocide had not occurred or that Armenians had committed genocide against Turks.
posted by bukvich at 12:27 PM on February 23, 2011 [1 favorite]


MetaFilter: You are not nearly as clever as you think you are.
posted by rough ashlar at 12:27 PM on February 23, 2011


Would it be possible to create speech honeypots, like websites and twitter accounts designed to draw out autocommenters?

If you are worried about astroturfing, think about what a speech honeypot would do to the political discourse.

The next logical step would be to create online agents to probe websites and twitter accounts to determine if they are speech honeypots. Then the speech honeypots need capabilities to determine if they are being probed...etc.
posted by kuatto at 12:29 PM on February 23, 2011 [1 favorite]


It's hard to imagine what the Air Force could use this for that wouldn't be a violation of the Hatch Act.
posted by Phssthpok at 12:30 PM on February 23, 2011 [2 favorites]


> Shades of Orwell's 1984.

It's a cookbook.
posted by hank at 12:30 PM on February 23, 2011 [6 favorites]


I've run into a Monsanto PR effort - make comments about Monsanto and bam! someone shows up to post pro-Monsanto stuff. (2001-2002 timeframe)

And the whole "Local City mom has 2 simple secrets for (white teeth/shrink belly/making $5000 a week)" webvertising effort is another example.
posted by rough ashlar at 12:32 PM on February 23, 2011


Then the speech honeypots need capabilities to determine if they are being probed...etc.

Turtles all the way down? Turtles all the way down.
posted by drezdn at 12:33 PM on February 23, 2011 [2 favorites]


People who scoff at the effectiveness of this haven't seen these cretins in action.

Before Digg completely overhauled their voting process (wrecking it in the process), they had serious trouble with the so-called Digg Patriots. A couple dozen conservative members, working solely within their own small collective (as far as we know) and using layman IP-masking techniques, were able to game the site. They gamed it so thoroughly that they were able to get rabidly biased and vitriolic posts from a not-even-one-week-old right-wing blog expressly created for the purpose of getting onto the front page of Digg onto the front page of Digg. That's just the most egregious example, but their efforts over the course of a few months poisoned discourse at the site to the point that even AskMe took note.

One of the most highly-trafficked websites in the world at the time, and the power to post to its front page was effectively exercised by a smallish collection of rank-and-file conservatives using the crudest of methods. If a bunch of disgruntled social media commenters can engineer that on their own, imagine what's possible with astroturf teams using these sophisticated sockpuppet tools with built-in backstories, automation, and the corporate or political backing necessary to operate a dozen instances per person full-time on a large scale.
posted by Rhaomi at 12:36 PM on February 23, 2011 [7 favorites]


So, this is just a PowerPoint deck to sell the services. Do we have any proof whatsoever that HBGary or anyone else has the technical ability to do all this without hundreds of human button-clickers?
posted by tippiedog at 12:47 PM on February 23, 2011 [1 favorite]


this without hundreds of human button-clickers?

Why without u want to do this? I have cousin's wife's brother running shop behind teh barber in South Ghaziabad, he'll make you a good price.
posted by infini at 12:49 PM on February 23, 2011


The Hatch Act exists to prevent senior government employees from getting too involved in domestic politics. It doesn't say anything about interfering in international politics....
posted by miyabo at 12:55 PM on February 23, 2011


'internet driving license'

GuyZero: I, for one, would fucking love this. I'll miss you all, but whatever. I enjoyed the internet a lot back in '91 as well.

Test 1: build a TCP/IP stack on a microcontroller. Extra points of you implement it in VHDL and a FPGA.

Test 2: There is no test 2. You just built an internets.


The problem is, robots are already building the internets. Trying to "humanize" the internet by making people do tasks that are now commonly completed by machines is bound to fail.

On the other hand: I say we make an "invite-only" internet. Build a parallel internet with IPv6, and everyone has a limited number of invites. If you invite a spammer, both the spammer and you get booted from the new internet.

You can get invited 3 times. If you get booted 3 times, you're free to play on the old internet, but the new one is off-limits to you. Make a fake account, you're booted, and so is your inviter. It'll be quiet at first, but so was the old internet. We'll need some Internet Police, and that'll require some collaboration and law-making, oversight, and whatnot, but it'll take some work to make something better for the god-awful mess this free-for-all has created.
posted by filthy light thief at 1:01 PM on February 23, 2011 [2 favorites]


I like the idea of imposing a syntax. Really, after having been a reader daily and for ages, even though I don't always post, it's pretty easy to tell a newbie from a seasoned MetaVet. Especially when in-jokes (bean plated as they may be) are used. When a thread really gets heated, anyone without chops can't stay around for long and is usually ignored or dismissed because of their poor argumentative and citation skills and is quickly booed off the stage. MeFi is one of the few places where I can see this strategy failing. Unless there are trolls hired SPECIFICALLY for the purpose of hanging around here.

On second thought, I guess there are volunteers who already do that...
posted by madred at 1:02 PM on February 23, 2011 [1 favorite]


If you invite a spammer, both the spammer and you get booted from the new internet.
You can get invited 3 times. If you get booted 3 times,


Quick! Flag that post as Spam so the thief can be removed from our sight here!

(And exactly HOW would this "U R A Spammer" brand be determined? Who'd watch the watchers?)
posted by rough ashlar at 1:07 PM on February 23, 2011


filthylightthief - I like the idea. The game rules sound like MeFi's Own - can't we just adapt them, why reinvent the wheel? I want somewhere to count beanses withoutses robotses.
posted by infini at 1:08 PM on February 23, 2011


I've always liked this CAPTCHA.
posted by exogenous at 1:15 PM on February 23, 2011 [23 favorites]


exogenous... exogenius link.
posted by madred at 1:21 PM on February 23, 2011 [1 favorite]


15 years after the bombs eliminated all life on earth, on a lonely website comments still come in.

GOPBoi759: Using atomic warfare is justified.
+79 upvotes
Teaparty4eva: You're so right
+30 upvotes
Imissjfk: Atomic war will kill us all
-101
posted by drezdn at 1:33 PM on February 23, 2011 [12 favorites]


As an honest crank, i'm appalled. My stupid, angry opinions are my own and if I repeat them constantly it's because I'm angry, not a bot. The products I shill are ones I love, so I'll mention them constantly.
I'd love to get paid for it though. I'm one if the few pro-development, pro-GM foods, anti-Green kids in Sydney's inner west. That's gotta be worth something to somebody...
posted by Lovecraft In Brooklyn at 1:37 PM on February 23, 2011 [2 favorites]


I bet anything the telecom companies have been doing this to badmouth and sew doubts about net neutrality.

You'd win that bet. One of my first more personal encounters with libertarian astroturfing happened after I blogged about net neutrality a few years back; like magic, a mysterious poster showed up in my quiet little blog to spread anti-neutrality propaganda. His contribution was just insensate corporate boilerplate, posted without any regard for what I had written.

I suppose that a major part of the trouble with this kind of nonsense--beyond the dissemination of bad ideas, the misrepresentation of support for bad ideas, and the clogging up of avenues for legitimate discussion--is the threat of widespread sciamachy. That is, that real people might try to engage libertarian astroturfers in good faith dialectic, thus wasting the time and energy of well-meaning people that might be used to do legitimate good in the world. Many of the suggestions here on how to manage and protect communities under assault from libertarian astroturfers are excellent ones, but I would add to them the spreading of the Monbiot articles far and wide, making sure that everybody who might be tempted to engage a stranger in debate online is aware that not only might they be addressing intractable fanatics, but possibly automated ideological noise generators. It's a damn shame that any real person should lose even a minute reading or responding to a fake persona; everyone should be informed of the possibility of encountering literally inhuman adversaries in the online public sphere.
posted by millions at 1:38 PM on February 23, 2011 [2 favorites]


MetaFilter's $5 entry fee is looking smarter and smarter.

Sure, until the robots steal my time machine!
posted by nickmark at 1:39 PM on February 23, 2011 [1 favorite]


To clarify, there isn't much application for captchas after the astroturfers already gain their accounts, but 1-captcha-answer-per-comment ala 4chan might prevent them from gaining too many accounts.
posted by jeffburdges at 1:49 PM on February 23, 2011


One other, equally plausible, reason that the military could be seekin pre-created IDs is for espionage. These days it is not enough to give an agent a pair of shades and a fake passport; if they lack a credible social media profile that tallies with their cover story, spies will be rumbled.
posted by rongorongo at 1:52 PM on February 23, 2011 [2 favorites]


A related and much more serious problem: the construction of real people who participate in web forums to advance a ruthless capitalist agenda without being paid because they sincerely believe it.

There are already methods to do this. I recently watched a talk by John Zeratsky, a User Experience Designer at YouTube, giving advice to Silicon Valley startups on exactly how to do this. This talk was originally given at Google's I/O developer conference. One recommendation was for companies to have personality - people don't like faceless corporations, if it's more human, personable and relatable, people are more likely to take their side and be more forgiving. The second was to engage in conversations with customers through twitter or blog posts - if customers feel engaged and involved with the company, and that their voice is important, they're more likely to identify with the company. He suggested that one important benefit of these techniques is that if your company is ever involved in a scandal - security breaches, leaking private customer data - these users will be the ones going to bat for you and defending you in forums where your name is being trashed. And because they're not actual employees of your company, they're perceived as more credible than a PR team.
posted by AlsoMike at 1:57 PM on February 23, 2011 [8 favorites]


Well, good, looks like the provenance of THIS account is established now. On to the next.
posted by Samizdata at 2:04 PM on February 23, 2011 [1 favorite]


To clarify, there isn't much application for captchas after the astroturfers already gain their accounts, but 1-captcha-answer-per-comment ala 4chan might prevent them from gaining too many accounts.

It's relatively easy to automate captcha entry if you have a decent number of people who will fill them out for you. 4chan themselves submitted over 200,000 captcha votes in a relatively short period of time during the whole marblecake thing. If you're only talking about establishing a minimal posting background for astroturf accounts you would not need a whole lot of human work to get it done.
posted by burnmp3s at 2:11 PM on February 23, 2011


Now, having been part of things like the whole Bieber-North Korea prank, and a user of things like TrackMeNot, a part of me deeply loves the idea of armies of advanced sockpuppets having huge, swirling memetic wars across the internet.
posted by Samizdata at 2:12 PM on February 23, 2011 [3 favorites]


a part of me deeply loves the idea of armies of advanced sockpuppets having huge, swirling memetic wars across the internet.

maybe not so much if your little grass hut happens to be in the way of said advance... the marching morons can't be stopped by mere human flies...
posted by infini at 2:28 PM on February 23, 2011


Somehow we need to adapt this approach to the intertubes. (Joel Watson at hijinksensue.com)
posted by Twang at 2:49 PM on February 23, 2011


Just imagine how bad things are going to get when voice synthesis gets really good. Already spammers are using callcenters in other countries to spam and scam people. What happens when that can be done automatically, with a computer trying to scam you?

Should be interesting to see how things go.
posted by delmoi at 2:50 PM on February 23, 2011


It's going to be like the post-apocalypse episodes from Dollhouse, only instead of reprogrammed people roaming the streets causing harm, it will be social network sockpuppet accounts, wandering cyberspace promoting products.
posted by hippybear at 2:56 PM on February 23, 2011


$5 an account starts to look more like an investment than a hurdle to jump over

An investment in cortex® banhammer polish.
posted by clearly at 3:00 PM on February 23, 2011


> Just imagine how bad things are going to get when voice synthesis gets really good. Already spammers are using callcenters in other countries to spam and scam people. What happens when that can be done automatically, with a computer trying to scam you?

One day in the not-too-distant future we're going to reach a singularity wherein we'll be forced to assume that everything we read, see or hear via any media delivery technology is some sort of advertisement.
posted by The Card Cheat at 3:34 PM on February 23, 2011 [3 favorites]


One day in the not-too-distant future we're going to reach a singularity wherein we'll be forced to assume that everything we read, see or hear via any media delivery technology is some sort of advertisement.

Sounds great. Might you have a link to such a technology? We are all interested in buying one.

phentermine horses hammers lobsterbisque moscow discount
posted by Inspector.Gadget at 3:51 PM on February 23, 2011 [5 favorites]



It's all a ploy by Matt and his minions to raise our $5 fee.
posted by notreally at 4:22 PM on February 23, 2011


These days it is not enough to give an agent a pair of shades and a fake passport; if they lack a credible social media profile that tallies with their cover story, spies will be rumbled

That may be a bit of an overstatement, it's not unheard of for someone not to have a Facebook or Myspace page and I doubt it would arouse suspicion except for perhaps people in their 20s or those who supposedly work in sales to the general public. It could certainly help create a backstory, but it wouldn't be necessary. I wouldn't imagine that it's essential if you're pretending to be someone who can be trusted with access for sensitive information--it might even be commonplace for such people not to give away too much information publicly through social media. I dunno, personally I don't even have a facebook page that uses my real name and I generally sign up for things using an alias and throwaway email accounts. Doesn't make me a spy.
posted by Hoopo at 4:33 PM on February 23, 2011 [1 favorite]


madred: "Especially when in-jokes (bean plated as they may be) are used. When a thread really gets heated, anyone without chops can't stay around for long..."

I see what you did there.
posted by mindless progress at 6:03 PM on February 23, 2011 [1 favorite]


I've run into a Monsanto PR effort - make comments about Monsanto and bam! someone shows up to post pro-Monsanto stuff.

I might have seen that here awhile back, when I mentioned the horrific problems I had with aspartame, and later MSG once the aspartame fucked me up. (as in, they were seriously thinking I might have multiple sclerosis.) Someone showed up and insisted that I (and my doctors) couldn't possibly be right, that I was full of shit. He actually went to the point of saying that it was impossible, and was convincing enough to even get a bunch of favorites, simply from speaking in a Very Authoritative Tone about medical science, which is full of uncertainty. People with actual medical expertise never talk that way, but he sure seemed to convince a bunch of people.

Totally felt like astroturfing to me.... someone with a paid opinion.
posted by Malor at 6:24 PM on February 23, 2011


and was convincing enough to even get a bunch of favorites

Uh... probably because they were right? snopes says that Aspartame doesn't cause health problems (With links to the FDA, Lancet and others). Wikipedia has a page on the 'Aspartame controversy'

Under the health and safty section they have this to say:
The safety of aspartame has been studied extensively since its discovery with research that includes animal studies, clinical and epidemiological research, and post-marketing surveillance,[51] with aspartame being one of the most rigorously tested food ingredients to date.[52] Peer-reviewed comprehensive review articles and independent reviews by governmental regulatory bodies have analyzed the published research on the safety of aspartame and have found aspartame is safe for consumption at current levels.[7][51][22][53] Aspartame has been deemed safe for human consumption by over 100 regulatory agencies in their respective countries,[53] including the UK Food Standards Agency,[54] the European Food Safety Authority (EFSA)[55] and Canada's Health Canada.[56]
MSG is a salt form of Glutamic acid, which your body produces on it's own, and is one of the key amino acids used to construct proteins. It glutamic acid occurs naturally in lots of foods. Here's what wikipedia says about sensitivity to it:
Subsequent research found that, while large doses of MSG given without food may elicit more symptoms than a placebo in individuals who believe that they react adversely to MSG, the frequency of the responses was low and the responses reported were inconsistent, not reproducible, and not observed when MSG was given with food.[18] No statistical association has been demonstrated under controlled conditions, even in studies with people convinced that they are sensitive to it.[18][19][20][21] Adequately controlling for experimental bias includes a double-blind placebo-controlled (DBPC) experimental design and the application in capsules because of the strong and unique after-taste of glutamates.[19] In a study performed by Tarasoff and Kelly (1993) 71 fasting participants were given 5 g of MSG and then administered a standard breakfast. There was only one reaction, and it was to the placebo in a self-identified MSG sensitive individual.[16] In a different study done by Geha et al. (2000), they tested the reaction of 130 subjects that reported sensitivity to MSG. Multiple DBPC trials were performed and only subjects with at least two symptoms proceeded. Only 2 people out of the whole study responded in all four challenges. Because of this low prevalence, the researchers concluded that the response to MSG was not reproducible
posted by delmoi at 6:52 PM on February 23, 2011 [8 favorites]


A related and much more serious problem: the construction of real people who participate in web forums to advance a ruthless capitalist agenda without being paid because they sincerely believe it.

How is that a problem? Thats real people saying they actually like / defend a company. Thats a real thing. If companies are behaving in ways that people like, it seems like a good thing.

Astroturfing is "bad" because people are paid and thus the motivations for their statements are hidden (like a reviewer not mentioning free stuff / investments / endorsements).

What you're describing is like Yelp reviews.
posted by wildcrdj at 6:54 PM on February 23, 2011 [2 favorites]


mindless progress: now for the apple sauce...
posted by madred at 8:55 PM on February 23, 2011


How is that a problem? Thats real people saying they actually like / defend a company.

When self-identity is constituted through systems of domination, you become attached to them and willingly participate in your own oppression. Let's not forget Adorno and Horkheimer too quickly, they can still teach us something: "The misplaced love of the common people for the wrong which is done them is a greater force than the cunning of the authorities."
posted by AlsoMike at 11:03 PM on February 23, 2011


That's cool, but so is this iPod. Besides, if more people know about my favorite band/author/game company/product they're more likely to get sales and stay in business
posted by Lovecraft In Brooklyn at 11:55 PM on February 23, 2011


As an honest crank, i'm appalled. My stupid, angry opinions are my own and if I repeat them constantly it's because I'm angry, not a bot

I think this is the fundamental problem with a lot of astroturf defense mechanisms, whether technical or social. It's really seductively easy to decide that someone you don't want to agree with is acting in bad faith, either trolling or a paid shill or whatever. It's also really easy to use the mechanisms for personal gain (does one of my co-revolutionaries think I shouldn't be made despot for life after the revolution? Then clearly he's in the pay of the opposition, take him out and shoot him). An even slightly overactive astroturf defense leaves you living in an echo chamber. If you're willing to accept living in an echo chamber, there are much easier ways to achieve it.

Malor / delmoi's aspartame thing, above, is a perfect example. Substitute creationism/evolution, or vaccination/antivaxxers, or whatever. If delmoi had sparse priors and didn't sound like a familiar mefite, it would be really easy to dismiss him as a shill. (Note that this analogy doesn't imply whether I think Malor or delmoi is right— just that they're both arguing in good faith, which I think they are.)

And because they're not actual employees of your company, they're perceived as more credible than a PR team

Oh come on, we happily dismiss people who say nice or bad things about Microsoft, Apple, or Linux, without even pretending that we think they're shills.

On the other hand: I say we make an "invite-only" internet.

Man, I haven't thought about USENET II in ages.

From their traffic stats, it doesn't look like anyone else has, either.

a part of me deeply loves the idea of armies of advanced sockpuppets having huge, swirling memetic wars across the internet

I love this idea too. It's all machine-wars-dystopia-y! It gained a lot of popularity in the context of spam as well, back in the day. Plenty of predictions that the "human Internet" would become a narrower and narrower slice of a network most of whose bandwidth was taken up by automatic spamming and spamfighting measures. The difficulties of spam fighting have a lot in common with the difficulties of astroturf defense, I think. Especially on proprietary systems like Facebook where you can't get a high-level picture of traffic to detect astroturf patterns more easily (well, Facebook itself could, but it's not in their interest to do so, not today).
posted by hattifattener at 12:55 AM on February 24, 2011 [1 favorite]


One day in the not-too-distant future we're going to reach a singularity wherein we'll be forced to assume that everything we read, see or hear via any media delivery technology is some sort of advertisement.

Gibson must be so proud.
posted by infini at 2:40 AM on February 24, 2011


These days it is not enough to give an agent a pair of shades and a fake passport; if they lack a credible social media profile that tallies with their cover story, spies will be rumbled

>>That may be a bit of an overstatement

Possibly - I am no expert. But my inspiration was this Economist article on "spycraft" which was written by people who may know a little more. It compares the two sides of espionage: the traditional one were agents are sent into the field with the new electronic one. They claim that as digital technology makes life ever easier for the latter group -search capabilities harder for the former. I would expect an agency intending to send a spy into the field would probably take the effort to hand-craft a social media profile. But for them to do this they would need to provide "friends" on Facebook/ LinkedIn etc. The accounts for those mythical contacts are the ones which might possibly be robo-astro-turfed.
posted by rongorongo at 4:47 AM on February 24, 2011


But for them to do this they would need to provide "friends" on Facebook/ LinkedIn etc. The accounts for those mythical contacts are the ones which might possibly be robo-astro-turfed.

How, then, do you tell real people with a real network apart anymore? Does this mean that a carefully curated and gate kept LinkedIn profile demonstrating a global network in one's field becomes meaningless nonsense rather than a web of reference for your work and reputation? I mean, I've got some guy with the job title POTUS in my network, how does that reflect upon me in today's world?
posted by infini at 5:04 AM on February 24, 2011


*looks in mirror*

*lifts hand to reach out and touch to see if 'for real'*

Kafka's current day question would be to wonder if he was a robot...
posted by infini at 5:05 AM on February 24, 2011


Which leads me to my next thought, that this rise of astro-robotic-sockpuppetry isn't just making conversations meaningless, its making your entire body of work (if there was a significant part of it online) meaningless. I could just as well write utter and absolute nonsense and it wouldn't matter anymore. There's not only no more credibility of information but none left for the people behind it either.

There'll be nothing left to watch, so what then would you sell, market or advertise against?
posted by infini at 5:14 AM on February 24, 2011


Which leads me to my next thought, that this rise of astro-robotic-sockpuppetry isn't just making conversations meaningless, its making your entire body of work (if there was a significant part of it online) meaningless. I could just as well write utter and absolute nonsense and it wouldn't matter anymore. There's not only no more credibility of information but none left for the people behind it either.

lolwut? If the stuff you're writing is not meaningfully distinguishable from autogenerated robot filler, you might wanna try a writer's workshop or something.
posted by nasreddin at 7:36 AM on February 24, 2011


If the stuff you're writing is not meaningfully distinguishable from autogenerated robot filler, you might wanna try a writer's workshop or something.

The other scenario is that key word searches will bring up all the lolwhutz thus threadshitting all over the quality comments, no?
posted by infini at 10:45 AM on February 24, 2011


I have no idea what you are saying.
posted by nasreddin at 11:00 AM on February 24, 2011


It's true.
posted by Burhanistan at 11:08 AM on February 24, 2011


« Older ET Sequel: "ET-X" [SLYT]...  |  Obama Justice Department Finds... Newer »


This thread has been archived and is closed to new comments