top | item 25630854

An automated pipeline for the discovery of conspiracy theories

114 points| iNic | 5 years ago |journals.plos.org | reply

109 comments

order
[+] peter_l_downs|5 years ago|reply
Linked site won't load so here's a vaguely-related book recommendation: Foucault's Pendulum by Umberto Eco. I don't want to spoil it, but it involves a group of writers who try to come up with a single, maximally compelling, conspiracy theory. Reading it sets up a super interesting tension -- the book is clearly fiction, and yet the conspiracy theory it presents is as compelling as any (and in fact incorporates many) of those you'll read "in the wild".

https://en.wikipedia.org/wiki/Foucault%27s_Pendulum

[+] KineticLensman|5 years ago|reply
I read Foucault's Pendulum and finished it, but it was heavy going. I then read "The Illuminatus! Trilogy" [0] which I enjoyed much more, because it was more intentionally cranky and funny. It is to conspiracy novels as "Airplane!" was to disaster movies, and then some.

[Edit] Both of them make Dan Brown look really, really unimpressive. If "The Da Vinci Code" is a vaguely thought-through concept aimed at general readers, "The Illuminatus! Trilogy" is an explosion in a conspiracy theory factory.

[0] https://en.wikipedia.org/wiki/The_Illuminatus!_Trilogy

[+] sdoering|5 years ago|reply
I can only agree with the recommendation. One fun fact I wanted to add is that in one of his writings Eco told the story of a letter he received because of said book.

In the book there is the description of a fire in Paris on a specific date. The reader wrote to Eco, that he must have gotten the date wrong, as he was at that place that night and that there was no fire.

Eco uses this as an example, that some readers do not understand the signifiers of fictionality and that authors sometimes have to deal with readers taking their fictional work literally.

[+] EamonnMR|5 years ago|reply
Second the recommendation, great book.
[+] ppod|5 years ago|reply
You'd think considering academics do the reviewing, proofing, and typesetting for free the $1600 publication fee might enable the publisher to at least make a site that can stand up to HN traffic. Here's an arxiv version: https://arxiv.org/abs/2008.09961
[+] RobertoG|5 years ago|reply
I have observed that speed of change in supporting facts is very important in Internet conspiracy theories (CT) vs. the old conspiracy theories.

It goes like this: you find a surprising fact that, if true, will support the CT. You go to investigate it, which, of course, take some time. You realize that the fact has a normal explanation but, when you go back to conversation of the CT, that's not a relevant fact in the conversation anymore. There is a new fact, or more, supporting the CT. Start the process again. It's impossible to catch with the fact that it's, at the moment, supporting the CT.

I think, this is different from the old way conspiracy theories, where people just believed highly improbable things.

It also makes it more appealing to more people, because, if you take it all globally and don't look carefully, it really looks like there is a lot of evidence.

[+] PeterisP|5 years ago|reply
Modern CTs are also sometimes the result of disinformation campaigns where (contrary to your example) there is not an attempt to provide multiple facts that would support "the" CT but rather continuously providing multiple plausible different (and incompatible) CTs to drown out the true situation - especially since most parts of the true explanation (which the disinformation campaign wants to muddle) would also overlap with one or more of the debunkable and debunked CTs, thus eventually getting to the desired end position of "ah, noone can really know what's true".
[+] jjk166|5 years ago|reply
I think the opposite is really the case.

Back in the day, fact checking was incredibly difficult. If you were a domain expert, you might notice something was off, then go to the library to double check the information, and if the library happened to have what you needed you could try to convince your close friends and family of the truth. The other 99.99% of what you heard on the news or read in the paper was simply accepted as fact. Conspiracy theories were, at the time, either what the news said were conspiracy theories, or ideas so fringe that the news didn't even talk about them.

Nowadays, we are a few thumb movements away from fact checking literally any piece of human knowledge. There hasn't been a sudden proliferation of fake news, it's just more readily apparent. As it has become easier to spot the holes in shoddy reporting, faith in journalistic institutions has plummeted. More people are getting news from alternative sources whose quality is variable. For better or for worse, there are simply more versions of stories nowadays.

To try to retain viewers, traditional media outlets have tried to cultivate an image of themselves as the arbiters of truth which they once defacto were, and to do so have been quite liberal with labelling their competition as conspiracy theories. We haven't seen any uptick in the number of people wearing tinfoil hats to keep out the mind control rays or claiming any given senator is a lizard person. Instead, we live in a world where you're labelled a conspiracy theorist if you don't think there is a shadowy cabal of government agents spreading misinformation to manipulate us or depending on which of the past two presidential elections you think was illegitimate.

Even small deviations from the narrative of any particular bubble are heretical, but that narrative, and every other one, is composed from the incomplete knowledge of fallible people, so invariably as time goes on you will start to notice inconsistencies in the story. In years gone by, we would have simply chalked this up to someone being misinformed, people could admit they were wrong and papers could print retractions and we'd all forget about it. But now the combination of the record being so readily accessible and the increased role of our media consumption in our personal identity means we have forgone nuance and are strongly pressured to adhere to increasingly absurd stories. It's so much easier to label the other side as stupid and crazy than to try to demonstrate the veracity of our position and face the fact that we are not entirely correct either, but if we do not strive for truth ourselves, it becomes easy for them to see the faults in our logic and further convince themselves that we, in fact, are the crazy and stupid ones.

tl;dr there are many more non-mainstream narratives being labelled as conspiracy theories at the same time that the mainstream narrators who would normally dispel them are losing credibility.

[+] haunter|5 years ago|reply
I'd much love to read a paper like this applied to conspiracy theories that became true (Epstein, PRISM etc). Basically reverse engineering
[+] joshdick|5 years ago|reply
The paper does analyze a real conspiracy: Bridgegate.
[+] pdabbadabba|5 years ago|reply
Agreed! But did I miss the part where the Epstein conspiracy theory was demonstrated to be true?
[+] djsumdog|5 years ago|reply
COINTELPRO, Operation Mockingbird, Bay of Pigs, the Sept 11 1973 CIA backed coup in Chile, Iranian-Contras, non-existant Weapons of Mass Destruction/Iraq, MKUltra, Tuskegee Syphilis Study, Church Committee ... the list is as long as you want to make it.
[+] m12k|5 years ago|reply
A headline that truly captures the times we live in. There's something poetic about using machine intelligence to detect human stupidity.
[+] throwaway80332|5 years ago|reply
I can't wait until this technology is perfected so that we can finally put an end to online disinformation for good.

Those dangerous idiots actually believe that a billionare pedophile spent decades inviting high-ranking politicians and influential decision makers to his private island for orgies with underage girls, and that when he was arrested he managed to hang himself while on suicide watch in a maximum security prison! They really are that deluded and dangerous. This needs to end, now.

The nutcases also believe that the governments are working with big tech to eliminate cash so that anybody who doesn't toe the line can be "permanently cancelled". Have you ever heard anything so ridiculous? I hope in the future we can just have an algorithm that disables the bank account of anybody who spews such divisive nonsense online. There is no place for that kind of hate in a society that strives for any kind of progress.

[+] jimktrains2|5 years ago|reply
I think it's counter-productive to label everyone who believes conspiracy theories "stupid". Not only does it make it more difficult to talk to and help change their opinion, it further polarizes use. Sure, I will readily admit that some people are just being obstinate and stupid, but when you have the places you get your news from feeding you misinformation, or worse, feeding you facts, but not all the facts, or the details get lost in transit (e.g. the cdc not recommending masks and then recommending masks), it's not hard to see how some things become ideas.

We can't help correct misconceptions if we believe they're driven solely by stupidity.

[+] noja|5 years ago|reply
For anyone looking for a way of talking to friends or family who are into conspiracy theories, this book is good: https://www.amazon.com/Escaping-Rabbit-Hole-Conspiracy-Theor...
[+] sandworm101|5 years ago|reply
Is there a book on how to not talk about conspiracy theories? I have some formerly right leaning elderly relatives who are now just plain crazy. They won't stop talking about ridiculous conspiracies. I don't open emails from them anymore. I've given up trying and wish they would at least stop talking to me about them.
[+] hammock|5 years ago|reply
>We show how the Pizzagate framework relies on the conspiracy theorists’ interpretation of “hidden knowledge” to link otherwise unlinked domains of human interaction, and hypothesize that this multi-domain focus is an important feature of conspiracy theories. We contrast this to the single domain focus of an actual conspiracy. While Pizzagate relies on the alignment of multiple domains, Bridgegate remains firmly rooted in the single domain of New Jersey politics

This study is interesting work to be sure, but the cause behind their hypothesis might be rooted in how journalism works, rather than whether it's truthful/factual or not.

Papers have a NJ politics "beat", with an editor and reporters who are experts in that single domain and are expected to produce work rooted in that specific domain.

Or worse, the "single domain" in some instances might be appearing because it was a story fed to the paper by a three-letter agency.

Citizen journalists (or whatever less charitable term you want to use) don't really have that. They can focus on whatever they want, and it's much easier to build a narrative that spans multiple domains.

[+] germinalphrase|5 years ago|reply
“ Papers have a NJ politics "beat", with an editor and reporters who are experts in that single domain and are expected to produce work rooted in that specific domain.”

I suspect that this is increasingly less true every year as local reporting is actively dying.

[+] boredumb|5 years ago|reply
We're actually doing a Rust application that downloads the image snopes is displaying for that particular conspiracy theory in order to feed them into a tensor flow model and predict the validity of future conspiracies.
[+] jackric|5 years ago|reply
Shut up and take my money
[+] throwaway91774|5 years ago|reply
I watched pizzagate grow from day 0, AMA.

I spent an inordinate amount of time on /pol/ around the 2016 election and when Wikileaks made the emails available I was among those hitting the random button to see what we could find. The authors of the paper did a great job in understanding a key feature; (vacuous) conspiraciy theories involve a lot of jumping to conclusions, and are all over the place (multi-domain). At the time there was so much momentum to find dirt on Hillary, that every silly comment seen through chan culture lenses became a zero-point energy engine, and suddenly we were swimming in mspaint.exe infographics pulled out of thin air.

I was never a believer though, but I did enjoy the shitstorm, and human trafficking is a real thing. I was disappointed that people was so fed up with the matter that Epstein's suicide was not a bigger deal with the general public.

[+] motohagiography|5 years ago|reply
Does nobody else appreciate the irony of using a graph, which essentially reduces to a conspiracy chart, to describe conspiracy theories?

CT's are just folktales and explanations for things people don't understand or control. A conspiracy theory only becomes dangerous or harmful when it threatens to upset an established order, which makes it oddly self fulfilling, since its real purpose is to facilitate organizing people around counter-establishment narratives. When you look at this paper as a new way to use technology to automatically detect counter-establishment narratives, it seems like pretty standard playbook for a secret elite coordinating to secure and expand their powers, which is hilarious, to me anyway.

However, the conclusion includes criteria for detecting actual conspiracies as separate from theories, "We hypothesize that three features—a single domain of interaction, a robustness to deletions of nodes and relationships, and a proliferation of peripheral actants and relationships—are key characteristics of an actual conspiracy and may be helpful in distinguishing actual conspiracies from conspiracy theories. "

I think they buried the lede on that one, as a heuristic for evaluating CT's and accusations of them could be super valuable.

[+] throwaway91774|5 years ago|reply
I'd say it's fitting, not ironic. Conspiracies do exist, and implicit conspiracy-like behaviour does emerge where minds converge too.
[+] aritmo|5 years ago|reply
Do these conspiracy theories appear organically or is there a special nudge to make them stick?

Who nudges the conspiracy theories to stick?

[+] crispyambulance|5 years ago|reply
It's a wild combination, everything from bored Estonian teens getting money from clickbait to state-sponsored disinformation campaigns with specific intentions.

The best writing about this IMHO is from Renee Diresta. She has papers, articles, talks, testified before congress, and founded the Internet Observatory project.

Diresta co-wrote a very comprehensive report on the activities of the Russian "Internet Research Agency": http://www.reneediresta.com/ira-report-4e8d0ff684.pdf It explains in gory detail how this "stuff" works when there are state actors involved.

[+] rsynnott|5 years ago|reply
I mean, they probably show up organically, but clearly sometimes people who should know better stoke the flames and spread them.
[+] jberryman|5 years ago|reply
The "Reply All" podcast episode #166 ("Country of Liars"), has a pretty good analysis of the history of the Q-anon conspiracy theory. fwiw
[+] throwaway91774|5 years ago|reply
I saw a lot of crap appear organically, but my guess is that by volume most is exagerated, clickbait, racebaitting stuff done as a dayjob, with wildly varying levels of sincerity. The few talented peddlers that create organic looking content make a huge difference, and some almost believe their own bullshit.

For reference: IMO, Alex Jones knew from day one that pizzagate was just bullshit made up on the internet, and kept distance accordingly. But there was so much intersection with his audience that he treaded carefully and did an obscure video on the side where he said he was just playing it cool to avoid persecution. Of course, if anything real came out of that, he would have claimed he was in on it from the start.

[+] Item_Boring|5 years ago|reply
I can’t answer your question but maybe you’ll find this paper interesting [0]. TLDR: social bots are highly responsible for spreading misinformation - and for such also conspiracy theories. By tagging people with a lot of followers and tweeting the information multiple times they attempt to make it go viral. Keep in mind that this paper concerns the 2016 elections.

[0] http://cs.furman.edu/~tallen/csc271/source/viralBot.pdf

[+] Tycho|5 years ago|reply
The important thing about 'conspiracy theories'® is that you have charlatans trying to profit off them by misleading people. What I mean is that if they come across information that hurts their narrative, they won't share it, while continuing to posture as honest investigators in search of the truth. And for the readers it's hard to detect this fundamental dishonesty (how do you know what you haven't been told?).

This isn't to say that there are no conspiracies, and people aren't justified in looking for explanations, but this mechanism of deceit is what makes it a controversial/goofball topic.

[+] guilhas|5 years ago|reply
I only believe in truths from my government or approved media sources. And I try to think as less as possível so I don't get too tired while doing the 9-5.
[+] arminiusreturns|5 years ago|reply
The important bits of the abstract to me.

> how conspiracy theories ... and their factual counterpart conspiracies... We show how the Pizzagate framework relies on the conspiracy theorists’ interpretation of “hidden knowledge” to link otherwise unlinked domains of human interaction, and hypothesize that this multi-domain focus is an important feature of conspiracy theories. We contrast this to the single domain focus of an actual conspiracy. By highlighting the structural differences between the two narrative frameworks, our approach could be used by private and public analysts to help distinguish between conspiracy theories and conspiracies.

I'm not sure how to interpret this new classification of true conspiracies as just "conspiracies" and false ones as "conspiracy theories" that they seem to be pushing in this paper. Many things others called conspiracy theory were eventually proven to be true, at least in some degree, and those things are often still called conspiracy theories because they have uncomfortable truths that people would rather not acknowledge as factual. This seems like an abuse of terminology in some way to me.

Before I delve to deeply into this, I will first say, however; As an open conspiracy theorist, (doing my best to "take back" the phrase), who tries to stick to the facts as much as possible, I have long thought about how a scientific approach could be used to prove the likelyhood of what others call conspiracy theories, and the longer I have thought on the matter the more I have great hope that some researchers or other would stumble upon this likely-career hurting approach to said theories. While I disagree with the papers characterization of certain theories as true and certain ones as false, both due to the black and white label and the lack of context primarily surrounding the accusations of the narrative framework of the true to be one that is in more flux and single domain vs the one of the untrue being constantly in less flux and multi-domain, I think this approach could be modified and put to better use beyond the limitations of this paper itself, primarily because the root of their study is the same as a serious conspiracy theory researcher: "actants (people, places, things), relationships between actants, and a sequencing of these relationships".

It is only in the application of this approach to center on stories and social media that a series of methodological mistakes emerge to weaken the paper. I could go into some of the details if wanted about these weaknesses, but in general on HN I try to keep the discussion more meta on the topic of conspiracy theories in order to not devolve the conversation too much. One example I will give however, is their overreliance on certain sources of data (twitter, reddit) that were considered at best secondary to the more deep and open conversations (the chans, irc, etc) being had on certain topics (pizzagate for example) juxtaposed against the reliance on UCLA aggregates of NJ newspapers on the topic of Bridgegate. I know, I was participating in all of the above when the last post which "broke the last straw" on reddit caused the sub to be banned. [1] Again, I don't say this to start a discussion on pizzagate as it is likely to devolve quickly, but rather to show I'm not making up my accusations of methodological weakness based on nothing. This sort of snowball methodological weakness then undermines their conclusion, which quickly goes off the rails into so many tropes and cliches not backed by their data and research it quite surprises me to see the the authors take a semi-defensible approach and allow it to be deteriorated by such a series of erroneous "conclusions". Go read the conclusion section for yourself if you think I am exaggerating.

1. https://archive.md/MrsGu

edit: in particular I would like to call out their overreliance on calling conspiracy theory discussion some variation of "imaginative interpretations of “hidden knowledge”". In reality, and this is something I frequently like to stress to my more logical/scientific friends who are skeptical of certain conspiracy theories, what many conspiracy theories rely on, often without being aware of it (to their detriment) is a series of inductive logical conclusions, as opposed to a series of deductive logical conclusions. Rightly so I say, because in the domain of conspiracy theory you often lack the hard evidence to back claims, and therefor must often rely on inductive logic instead (and the lack of deductive evidence does not alway indicate untruthfulness as is often assumed). This is one of the keys that helped me get past many issues, because when it comes to conspiracy theories true or not, the real crux is the probabalistic likelyhood of truth which can be modified as more data points emerge, not some black and white true and untrue label.

[+] webmaven|5 years ago|reply
> This is one of the keys that helped me get past many issues, because when it comes to conspiracy theories true or not, the real crux is the probabalistic likelyhood of truth which can be modified as more data points emerge, not some black and white true and untrue label.

I think part of the problem here is that most of what we call conspiracy theories that are true aren't actually conspiracies per-se, except in the loose sense of "a conspiracy of silence", and don't particularly intersect with things like secret societies. I mean, sure, there are efforts made to conceal things from the public, and yes, there is a certain amount of coordination among powerful actors where their vested interest align, what else would you expect to happen in the real world? For every amoral person to act as a solitary megalomaniacal villain?

But none of that particularly implies that this sort of thing is planned ahead or centrally organized in any particular way or there wouldn't be any need for in-person meetings like the Bilderberg Group (which didn't even meet in 2020). It certainly doesn't imply that such conspiracies deliberately interlock with each other except as you would expect simply from survivorship bias.

I mean, organized crime is a conspiracy. A large corporation evading responsibility for a chemical spill is a conspiracy. Cartels doing price fixing is a conspiracy. The US military flexing in support of private interests is a conspiracy. Lobbyists getting exceptions and loopholes enacted into law is a conspiracy. Politicians accepting bribes is a conspiracy.

So conspiracies as such aren't unknown. Such things come to light all the time in part because maintaining secrecy is so damn hard and there are many interests aligned in exposing them. We should encourage that where and when we can.