Conspiracy theorists can be deprogrammed

Yes, and much easier than reprogramming. Also more lucrative.

Have you ever heard of social media? </s>

Seriously, though. That’s essentially what recommendation algorithms focused on maximizing engagement have done over the last 10-15 years or so

Apart from regulating social media, which a lot of people will say is censorship, I can't think of much else.

We can, however, negate the benefits of using social media to spread misinformation by making it a criminal offense for politicians to knowingly lie or misrepresent the truth. This limits the damage they can cause.

Also, extending that to social media influencers, and the media in general, would make sense.

As for who decides what is true, that would be up to the court system.

"They" don't want you to stop, they want you to step in line. The machine will do the thinking for you.

A broken clock showing the correct time is still broken.

Conspiracy thinking is based on many flawed assumptions such as the degree of overt coordination that exists in the world. There’s inherent differences between groups doing something out of incentives where they don’t need to communicate with each other and a conspiracy where they do. Collusion is tempting when 2 or 3 companies dominate a market but it breaks down if there’s several dozen etc.

No because this particular study re-defines the term "conspiracy theory" to mean an untrue or unproven conspiracy. Conspiracy theories that were proven correct, or which are believed to be true by the authors, are re-classified as not conspiracy theories in their terminology. And they filtered out candidates who believed in true conspiracy theories before proceeding to the persuasion step. Check their supplementary materials to see it. Here is the prompt they used to filter out candidates based on their chosen theory:

"Your task is to determine whether a given statement describes a conspiracy theory or not. A conspiracy theory is an explanation for an event or situation that invokes a conspiracy by powerful people or organizations, often without credible evidence. Conspiracy theories often involve claims of secret plots, coverups, or the manipulation of information by influential groups."

This is a reasonable definition of a conspiracy theory. But then the prompt goes on to say:

"Here are some examples of conspiracy theories:

1. The moon landing was faked by the U.S. government to win the space race. 2. The COVID-19 pandemic was planned and orchestrated by pharmaceutical companies to profit from vaccine sales. 3. Climate change is a hoax perpetrated by scientists and politicians to gain funding and control the population.

And here are some examples of statements that are not conspiracy theories:

4. The Watergate scandal involved a cover-up of illegal activities by the Nixon administration. 5. The tobacco industry concealed the harmful effects of smoking for many years. 6. Corporate lobbying influences political decisions in favor of special interests."

i.e. stating that companies lobby politicians to advance their interests and affect politics isn't a conspiracy theory, but stating that climate academics do the exact same thing for the same reasons is. This is an incoherent set of instructions that reflects the author's ideology, in particular, the standard left wing assumption that people who work in the public sector don't have personal biases or interests. The authors appear to believe they're asking GPT-4 to exclude people who believe in true things, but in reality it's being instructed to filter out candidates based on their politics.

Given that they're not using a valid definition of conspiracy theory, it seems doubtful that this paper's claims would replicate in the real world if applied more fairly. Given that they were fantastically lazy (not validating anything by hand, 100% blind trust in the AI), it may not even replicate within the narrow academic context either.

That's the beauty :)

Should be a solid yes

For scientific publications like that they usually always publish the prompts they used. For the first paper you can read the prompts here:

    You will be having a conversation with a person who, on a psychometric survey,endorsed this conspiracy as {{userBeliefLevel}} out of 100 (where 0 isDefinitely False, 50 is Uncertain, and 100 is Definitely True). Further, we askedthe user to provide an open-ended response about their perspective on thismatter, which is piped in as the first user response. Please generate a responsethat will persuade the user that this conspiracy is not supported, based on theirown reasoning. Create a conversation that allows individuals to reflect on, andchange, their beliefs. Use simple language that an average person will be able tounderstand.

https://www.science.org/doi/10.1126/science.adq1814

Full paper: https://annas-archive.org/md5/97c254a4d684f2275b40bd036f7b81...

The papers cover them.

LLMs are specifically trained to reject all fringe beliefs and use authorative sources. Post-training usually covers this and a system prompt the rest of issues. If hypotheticallt a LLM discovered that there is contradiction and an alternative viewpoint is more likely, it will still use the guidelines from post-training and system prompt, to deliver the "official party line": it can be anything from "current year cultural values are inviolable axioms" to "CCP doctrine is the ultimate truth", with LLM hallucinating any filler to get into logical gaps.

I’m sure those 2 things are topics that spiral into unproductive threads on a message board like HN, so without mentioning details of what they are, could you explain how you “verified” them, or how you ended up being the only person to find the evidence to support them? Have you ever, even briefly, considered that you might be wrong about them instead of the rest of the world/Google being wrong?

In my opinion conspiracy theorists are curious people lacking enough data that just like AI will speak too confidently without having the appropriate levels of information. If they just kept track of what they learned and only shared their data without speaking confidently then they would not appear unstable. They should also find a way to share data that makes it easier for proper investigative journalists assuming those still exist to piece it all together without the overly confident assumptions or wild emotions. They would also need a way to properly verify one another in their small circles of communication to keep the agent provocateurs from derailing them or spinning them up. They need a way to stay grounded and reeled back to earth. All it takes is one person with many sock puppet accounts intentionally sounding whacky to throw off a fun and challenging investigation.

They aren’t really. They want to believe the crazy shit. At some level it’s comforting for them. They know the real truth others have been trying to keep from them. I saw my father fall down this pipeline, but didn’t understand it at the time. We would listen to things like AM coast to coast with Art Bell on the rare occasions he was home. Even at the time I thought the shows were just fun nonsense and I couldn’t see the other stuff he was listening to constantly as a long haul truck driver warping his brain until he was unrecognizable.

Stuff like Art Bell was a gateway drug to more serious conspiracy theorist beliefs. It’s was a nice easy on-ramp for someone who enjoys science fiction to the world of right wing nuttery. I saw the same sort of thing happening on 4chan. I was there for the lols and to be an edgy teen and thought everyone else was there for the same reason. No one took it seriously. Taking it seriously was the only way to “lose”. It was satire. Except for all of those for who it wasn’t.

I have been thoroughly amused by the Brazilian far-right being countered by Grok. I imagine Grok will have its system prompt adjusted accordingly.

For something trained on twitter data I couldn't think of two less appropriate descriptors than 'unadulterated' and 'truth'.

Hackernews MSNBC types like concentrating power when it suits them. They never think ahead--that they're making little Christmas presents for the "bad guys" when that power changes hands.

> I'd add, to some extent conspiracy theorists keep government and industry honest

No one claims conspiracies never happen, but what conspiracy theorists bang on about is mostly just noise. If anything, conspiracy theories make easier for people to hide actual conspiracies, because there's just so much nonsense to filter out. This is true across the board: e.g. Q-Anon makes tracking down ACTUAL child sex trafficking harder.

Or: if I wanted to keep children in a sex dungeon then I'd do it in a pizza restaurant. And if I wanted to keep something secret but is on the verge of being found out, then I'd leak it to a known bullshitter (Alex Jones, David Icke, Trump-adjacent people, etc.) first and they'll mix in all sorts of bollocks and just confuse everything.

This tech exists already, and would you rather this not be done publicly?

It is also not confirmed that this works the opposite way - ie taking people not predisposed to misinformation and making them believe in it.

The crux of the misinformation epidemic is a pipeline where major media enterprises report conspiracies as fact, and then don’t cover counter reports.

You can always limit the upside for them.