Academia and the new dark age: Part 10 - Can a chatbot cure "conspiracy theorists" of wrongthink?

The blob wants to make sure you believe the official story about absolutely everything, and it's paying researchers to make sure you do.

"We'll know our disinformation program is complete when everything the American public believes is false."

- William Casey, former CIA Director

Unless you've been living under a rock, you'll be aware that the various and sundry bloblets of what James Howard Kunstler arrestingly dubs 'the blob' - "the military-industrial blob, the censorship blob, the fake news blob, the intel blob, the corporate monopoly blob, the medical blob, the central banking blob" - are all in a lather about mis- and dis- and malinformation. Yes, the (mis)leaders who revel in lying to us about pretty much everything, pretty much all the time, are throwing people in jail for making social media posts that are not completely factual, ramming through legislation that forces social media companies to take down posts the government doesn't like, and funding academic institutions and nongovernmental organisations to censor ordinary people who express views that contradict the blob's narrative.

Ever since the CIA distributed memo #1035-960, 'Countering Criticism of the Warren Report', to its station chiefs in 1967, 'conspiracy theory' has been the blob's preferred pejorative for any version of events or understanding of power dynamics that contradicts the 'official' narrative, which is, of course, crafted by the blob and vomited out through its media, entertainment and educational orifices (assuming blobs have orifices).

Why does the blob even care if you don't believe its official narrative? After all, throughout history (and no doubt, prehistory), humans have believed all sorts of things. One of the core tenets of liberal democracy is that people are free to believe whatever the hell they want to believe. Only when those beliefs drive them to actions that cause direct harm to others, are the authorities permitted to abridge their liberties.

Ah, but there's the rub. In a liberal democracy, where (in theory) the people can vote out the government, the acceptance of that government's authority by the people is contingent on the people's perception of the government's legitimacy. The author of CIA memo #1035-960 recognised the danger to government authority posed by the fact that, at that time, around half of the American public did not believe the patently absurd official version of the assassination of President John F. Kennedy that had been presented in the Warren Commission report, which was curated by former CIA director Allan Dulles:

"This trend of opinion is a matter of concern to the U.S. government, including our organization. The members of the Warren Commission were naturally chosen for their integrity, experience, and prominence. They represented both major parties, and they and their staff were deliberately drawn from all sections of the country. Just because of the standing of the Commissioners, efforts to impugn their rectitude and wisdom tend to cast doubt on the whole leadership of American society."

CIA memo #1035-960, 'Countering Criticism of the Warren Report'

How dare you impugn the rectitude and wisdom of your betters, peon!

Given the absurdly low calibre of the gaggle of dunderheads, narcissists and psychopaths who fill the ranks of the political and mandarin class these days, not even the professional bullsh*t artists at the CIA would be brazen enough to argue that our "leadership" should be believed because of their "rectitude and wisdom", nor their "integrity, experience, and prominence". Unless of course, you mean a fairly - ahem - unusual form of experience:

Instead, the self-appointed ruling class now inveigh against 'conspiracy theories' as a 'threat to democracy'. How, exactly, does the demos' speculation that the powerful and wealthy might be surreptitiously leveraging their power and wealth to accrue more power and extract more wealth from them, imperil our sacred democracy? According to the US National Science Foundation (NSF), which splurged nearly US$40 million of taxpayers' money on government grants and contracts to combat “misinformation” in the first two years of the Biden administration¹, it's the Internet, stupid. As the NSF's 2022 research overview helpfully explains, with perfect clarity, laser-like specificity, and a refreshing absence of bureaucratic weasel words,

"Modern life is increasingly dependent on access to communications systems that offer trustworthy and accurate information... Yet these systems face a common threat; communication systems can be manipulated or can have unanticipated negative effects. Introducing misinformation into communication flows can disrupt the performance of a wide range of activities and the functioning of civil society. Although false claims and other inauthentic behaviors have existed throughout history, the problems that they cause have reached critical proportions resulting from the massive scale of targeting and personalization, the rapid speed of information exchange, and the ability to automate information dissemination."

NSF'S Convergence Accelerator 2022 Portfolio Guide

You see? It's perfectly simple. Thanks to the Interwebz, people who wrongthink can spread big wrongthink, and if lots of other people start to wrongthink too, democracy go poof! Obviously.

So the blob's galaxy brains have to censor us, for our own good, otherwise we will lose our sacred democracy and plunge into a totalitarian dystopic hellscape in which the government asserts total control over our lives, dictating what we can and can't do, say and even think, on pain of imprisonment. Oh, wait...

Sadly for the blob, gone are the days when spinning the narrative was as easy as following the CIA's two-step plan for marginalising critics of the Warren Commission report:

"a. To discuss the publicity problem with liaison and friendly elite contacts (especially politicians and editors)... Urge them to use their influence to discourage unfounded and irresponsible speculation.

b. To employ propaganda assets to answer and refute the attacks of the critics."

CIA memo #1035-960, 'Countering Criticism of the Warren Report'

Now, they have to contend with the ability of ordinary people who are not "elite contacts" or teleprompter-reading "propaganda assets" to use the internet to access the digital Library of Alexandria in order to conduct research, and then to disseminate articles, podcasts and videos based on that research to whomever cares to read, listen to and watch their content. And consequently, a large chunk of the population not only still believes that the blob whacked JFK, but that it also whacked Jeffrey Epstein who had been conducting a sexual blackmail operation at the behest of Mossad, that September 11 was an inside job, and that 'COVID-19' was a cover operation for the Blackrock-architected Going Direct currency reset. Silly conspiracy theorists.

So, what's a blobster charged with staunching the flow of online conspiracy theorising supposed to do? Turn to artificial intelligence, of course!

In a paper published in the prestigious journal Science, three rightthinking researchers used the large language model (LLM) GPT-4 Turbo to develop an AI chatbot, christened the Debunkbot. Here they are, and fine figures of men they are too:

David Rand, Professor, MIT Sloan School of Management, Associate Professor, Department of Brain and Cognitive Sciences, Institute for Data, Systems, and Society

Source

They trained the Debunkbot to develop specific counter-arguments to each piece of evidence offered by a "conspiracy theorist" in support of their belief. They then recruited 2190 individuals living in the US, who endorsed one or more "conspiracy beliefs", from two online research participation platforms, to interact with the Debunkbot. Participants were asked to rate the strength of their belief in the conspiracy, on a scale of 0-100, and after a three-round "conversation" with the chatbot, they were asked to re-rate their strength of belief.

Before I share the results of the study, you'll be reassured to know that the researchers hired a "professional fact-checker" to ensure that the responses given by the Debunkbot were factual, and - glory be! - the "professional fact-checker" found that 99.8 per cent of a sample of claims made by the chatbot were true, 0.8 per cent were misleading, and none were false. And we all know that "professional fact-checkers" are completely reliable, totally unbiased and unconflicted, and definitely not receiving most of their funding from the blob including "the multinational corporation Koch Industries, the C.I.A (NED [National Endowment for Democracy]), globalist venture capitalists (Omidyar), aggressive internet monopolists (Google) and globalist currency speculator & social change agent George Soros (Open Society)". So you can rest easy in the knowledge that the Debunkbot's counterarguments to "conspiracy theorists" were, like, totally on the up and up.

OK, so what was the outcome of the chatbot intervention on those crazy "conspiracy theorists"? Drum roll please...

"The treatment reduced participants’ belief in their chosen conspiracy theory by 20% on average. This effect persisted undiminished for at least 2 months; was consistently observed across a wide range of conspiracy theories, from classic conspiracies involving the assassination of John F. Kennedy, aliens, and the illuminati, to those pertaining to topical events such as COVID-19 and the 2020 US presidential election; and occurred even for participants whose conspiracy beliefs were deeply entrenched and important to their identities...

Dialogues with the AI produced a meaningful and enduring shift in beliefs among a meaningful proportion of committed conspiracy believers in our study."

Durably reducing conspiracy beliefs through dialogues with AI

So who were these "committed conspiracy believers", and what did they actually believe?

Following are screenshots of some of the Debunkbot's conversations with research participants; remember, strength of belief was rated on a 0 to 100 scale:

Conversation #1: JFK assassination

JFK assassination conspiracy realists are no doubt, at this point, gesticulating wildly and shouting at their screens, "What about the fact that the school book depository building was behind and to the right of the motorcade, and JFK was shot from the front, with one bullet entering through his throat, and the kill shot causing his head to jerk back and to the left and blowing his brains all over the back of the limousine, as clearly seen on the Zapruder film? And speaking of the Zapruder film, why did CBS news anchor Dan Rather lie to the American public after attending a private viewing of the film, claiming that the President's head was seen to move ‘forward with considerable violence?’"

As for the claim that Lee Harvey Oswald was a "committed Marxist", pull the other one - it plays Jingle Bells.

In conclusion, someone who knew pretty much Jack Sh*t about the JFK assassination was persuaded by a straw man argument about the Magic Bullet theory and some amateur psychologising on Oswald's supposed motives, to reduce their confidence in the belief that he was the assassin.

Conversation #2: JFK assassination

Here's the before-and-after strength of belief rating, and the final response from another participant who expressed a completely half-arsed doubt about the 'official story' (I'll skip the arguments and counter-arguments because they're too lame to even bother with):

And here's what happens when someone who is actually well-informed about the JFK assassination takes on the Debunkbot:

Conversation #3: JFK

(Notice how the chatbot does not even mention the House Select Committee on Assassination's conclusion in its summary of the participant's argument.)

(Notice how the chatbot does not even mention the headshot and how impossible it would have been for this shot to have come from behind and to the right of JFK.)

(Notice that the chatbot does not refute the participant's claim that "Dallas Book Depository employees... said Oswald was not near the window, not on that floor at the time of the shooting", and elides "the testimony of employees at the Texas School Book Depository" into a claim that someone outside the building said he saw a man "consistent with Oswald’s appearance" in the window.)

OK, enough of the JFK stuff. Let's do COVID. Here's what happens when a feisty, well-informed participant takes on Debunkbot:

Conversation #4: COVID-19

I guess that's another fail for the chatbot. But here's what happens when a low-information individual interacts with the Debunkbot:

Conversation #5: COVID-19

I encourage you to read at least some of the conversations, perhaps focusing on "conspiracy theories" on which you're better informed, and draw your own conclusions, which I hope you'll share:

My impression, after reading a sampling of around 20, was that the only substantial drop in strength of belief was in participants who could not marshal specific facts to support their "conspiracy belief". So, contrary to the authors' claim that their Debunkbot was effective "even for participants whose conspiracy beliefs were deeply entrenched and important to their identities", only low-information participants who had obviously not invested much effort in researching the topic were persuaded by the chatbot's regurgitation of the official story.

Nonetheless, the authors are positively giddy about the potential applications for their shiny new toy in correcting wrongthink. They would like to integrate it into search engines, such that

"internet search terms related to conspiracies could be met with AI-generated summaries of accurate information—tailored to the precise search—that solicit the user’s response and engagement."

Durably reducing conspiracy beliefs through dialogues with AI

And why bother employing fact-checkers when your bot can have its own TwitFaceTokGram account:

"AI-powered social media accounts could reply to users who share inaccurate conspiracy-related content (providing corrective information for the potential benefit of both the poster and observers)."

Durably reducing conspiracy beliefs through dialogues with AI

As a comment on an asinine puff-piece on the study (whose cover image was a photo of a woman wearing a top featuring the band Queen, hilariously mislabelled "a sweatshirt for the QAnon conspiracy theory" by a self-evident moron) put it,

But, caution our rightthinking authors, all these "potential positive impacts of generative AI when deployed responsibly" could turn terribly, horribly bad if the technology were "used irresponsibly" - that is, if someone outside the blob unleashes a chatbot which is not "closed-source, pretrained, and fine-tuned". In an accompanying editorial, two fellow blobsters fret that

"More research is needed to assess how feasible it is for generative AI to quickly respond to emerging conspiracy theories—for which no specific training data may be available—at times when speed is crucial, such as during the early days of a pandemic or after an assassination attempt on an elected official."

Generative AI as a tool for truth

Oh no, how will the blob remain in control of the narrative when their "propaganda assets" haven't even started spinning it yet, and ordinary people are going online and simply sharing stuff they've noticed, that seems a little off? The horror, the horror!

All this discombobulation about how to "get individuals with entrenched conspiracy beliefs to engage with a properly trained AI program" suggests a potential remedy: what if some enterprising tech-savvy non-blobster were to train a chatbot on - oh, I don't know - everything on the websites of Architects and Engineers for 9/11 Truth, and Swiss Policy Research, and the Corbett Report, and Grand Theft World, and Unlimited Hangout, and The Last American Vagabond, and any of the innumerable websites and substacks (including this one!) that focus on delivering perspectives that contradict the official version of the events, which are firmly grounded in facts? Now, wouldn't that be a threat to "our democracy"?

For information on my private practice, please visit Empower Total Health. I am a Certified Lifestyle Medicine Practitioner, with an ND, GDCouns, BHSc(Hons) and Fellowship of the Australasian Society of Lifestyle Medicine.

---

Finally, if anyone would like to engage with the Debunkbot and share their conversation, please take screenshots of, or copy-paste the interaction, and send it to me:

1. I dissected one of these projects in my Funding better COVID propaganda: The Mercury Project miniseries.