Essay by Eric Worrall
“… However, there are still numerous hurdles to overcome before the goal of automated debunking is achieved …”
Hierarchical machine learning models can identify stimuli of climate change misinformation on social media
Communications Earth & Environment volume 5, Article number: 436 (2024) Cite this article
Abstract
Misinformation about climate change poses a substantial threat to societal well-being, prompting the urgent need for effective mitigation strategies. However, the rapid proliferation of online misinformation on social media platforms outpaces the ability of fact-checkers to debunk false claims. Automated detection of climate change misinformation offers a promising solution. In this study, we address this gap by developing a two-step hierarchical model. The Augmented Computer Assisted Recognition of Denial and Skepticism (CARDS) model is specifically designed for categorising climate claims on Twitter. Furthermore, we apply the Augmented CARDS model to five million climate-themed tweets over a six-month period in 2022. We find that over half of contrarian climate claims on Twitter involve attacks on climate actors. Spikes in climate contrarianism coincide with one of four stimuli: political events, natural events, contrarian influencers, or convinced influencers. Implications for automated responses to climate misinformation are discussed.
Read more: https://www.nature.com/articles/s43247-024-01573-7
The section on the ultimate goal of the project is fascinating.
… These findings have practical implications. Adopting our model could help Twitter/X to augment and enhance ongoing manual fact-checking procedures by offering a computer-assisted procedure for finding the tweets most likely to contain climate misinformation. This adoption could make finding and responding to climate-related misinformation more efficient and help Twitter/X enforce policies to reduce false or misleading claims on the platform. Yet environmental groups have shown that Twitter/X ranks dead last among major social media platforms in its policies and procedures for responding to climate misinformation and there is little evidence that X will improve these procedures in the near term41. Alternatively, our model could provide the basis for an API that Twitter/X users could employ to assess climate-related claims they are seeing in their feeds. Overall, the potential practical applications of our model underscores the need for continued academic work to monitor misinformation on Twitter/X and raises important questions on the data needed to hold social media platforms accountable for the spread of false claims.
However, there are still numerous hurdles to overcome before the goal of automated debunking is achieved. An effective debunking requires both explanation of the relevant facts and exposing the misleading fallacies employed by the misinformation. Contrarian climate claims can contain a range of different fallacies, so automatic detection of logical fallacies is another necessary task that, used in concert with the CARDS model, could bring us closer to the ”holy grail of fact-checking”17. …
Read more: https://www.nature.com/articles/s43247-024-01573-7
Our old friend John Cook was involved with this paper. According to the paper he is now a researcher for the rather Orwellian sounding Melbourne Centre for Behaviour Change, which appears to be part of the University of Melbourne Psychology Department. The mission statement of the center is “Harnessing research and education to produce sustainable, durable changes in behaviours, policies, and practices that will enhance lives, livelihoods, and environments.“.
John Cook suffered a public embarrassment back in 2013, when his collection of creepy artwork became public knowledge, thanks to poor website hygiene. The artwork featured a self portrait of John Cook dressed in a NAZI uniform, and included a picture of prominent climate skeptics, including Anthony Watts, dressed as semi-naked Roman gladiators. But I’m sure this unusual art hobby in no way affects Cook’s work for the Centre for Behaviour Change’s mission to “produce sustainable, durable changes in behaviours“.
I couldn’t figure out who funded the AI paper, but given the close author affiliation with major Australian and British universities, I’m pretty confident tax dollars feature somewhere in the picture.
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
They used to say that one man’s terrorist is another man’s freedom-fighter. Should we now update this to “One man’s truth is another man’s disinformation”?
Yes since each side in the debate calls the other side’s arguments disinformation. The word has no actual meaning.
It’s all just “information”.
Why not let the individual have all the information available then determine for themselves?
You can’t do that !!!
People would soon come to politically incorrect/inconvenient conclusions & realize the government had been lying for years, making the population difficult to control & diminishing the power of the executive.
Control the language – control the ideas.
K. Marx
control the education of the youth
Could you reference the quote?
Because they might reach the ‘wrong’ unapproved conclusion.
And their response would be, ala Jim Carrey in ‘Liar, Liar,” “Because it’s DEVASTATING to my CASE!”
Misinformation, and its sibling disinformation (which is merely misinformation with implied knowledge of the supposed falsity of that information) is nothing more than the ruling body, usually a government, claiming to have a Divine Right to absolute truth. It is a claim of Godhood, of infallibility.
To deny the opposing view in any argument weakens all sides. As John Stuart Mill stated much more eloquently than I could in On Liberty:
Isn’t ’Disinformation’ just the alternative spelling for ‘Government’?
It’s a stupid sounding word.
Recently I asked an AI “If AI existed a hundred years ago would the radical and uniformly debunked theory of “Continental Drift” ever gotten any traction. At least it answered that no the theory would have been labelled misinformation.
Any paper with John Cook involved is automatically suspect. An honest AI would have already easily found the following:
Proving climate science IS unreliable.
And El Hierro proves that clean energy does NOT work.
Both being specific sets of facts Cook in graphic 1 claims AI says are not true.
Cook needs a better AI. His is too easily proven false. What else would we expect from a Cook paper in the well know journal Communications Earth and Environment, whose sole purpose is to communicate alarmist propaganda. Dressing it up as AI doesn’t work any more than Cook’s former website ‘Skeptical Science’ wasn’t skeptical.
Said differently, putting lipstick on a pig doesn’t change the pig.
Rud, nothing about the climate crisis is honest, but more importantly, that’s not the goal of this AI. Note all the comments about “helping” twitter. That’s code for “we can build a tool and then get government to shove it down X’s throat”.
Enforcing the alarmist narrative is all this tool is about.
As I say at my GelbspanFiles blog, he is that John Cook, thankfully no relation to me, whose PhD degree is in philosophy, where everything he knows about ‘industry-paid skeptic climate scientists’ was taught to him by Naomi Oreskes – not through her idiotic 2010 “Merchants of Doubt” book, but instead via her chapter contribution within another obscure 2010 book where she hurled that unsupportable accusation.
AI: rubbish in, rubbish out. Just another means to promote and misrepresent self-evident truths. Think about when Gemini was rolled-out it fell flat on its face. Not very impressive. But then again, it was a Google product. Now you know.
I use “Claude.ai” for checking and explaining things and I find it quite helpful. When it answers a question it usually gives a pretty good answer, reasons, and options where needed. I also try and reuse saved chats so it has some context. It’s much easier than searching by hand for lots of things.
Sorry, I am going to have to ask to see the receipts.
Claude.ia is a chatbot on top of a large language model. Not even Anthropic claim it is capable of reasoning. It is generating text that has the superficial appearance of an answer. That is all it does.
It may well be able to turn turgid academicky prose into concise English, and to that extent it would be useful. But it absolutely does not understand anything. Use it by all means, but unless you—you—keep your brain switched on throughout, it is no better than pounding keys on a calculator and thinking the number it shows is an answer just because it is a number.
(Anyone recalling my recent posts on this subject might be getting the idea that I think large language models are a bigger red-herring than climate models. You would not be wrong…)
More than a few companies have been brought down by an errant spreadsheet.
Billion dollar probes crashing into a planet’s surface instead of orbiting.
Highway overpasses off by 9 feet after a decimal dropped.
Refineries exploding.
That kind of stuff was pounded into my head in Engineering school and yet…
Such errors would be less common if whoever was responsible was held responsible.
“Each accident has a name, surname and a position”
Lazar Kaganovich
I’ve read- may or may not be true- that back in the ancient world, when the lintel was set in an arch, the builder would have to stand under it- good motivation to get it right.
Agreed. I wish everyone would quit calling these simulations AI. It is not AI just a program to simulate AI. In a real AI scenario the computer programs itself. 🤷♂️😒
It IS “AI” as long as you understand that “AI” stands for “Automated Idiocy.”
On that I’m in 100% agreement. 😃
The debate is in fact a tree structure, the issue tree i discovered just over 50 years ago. See http://www.stemed.info/reports/Wojick_Issue_Analysis_txt.pdf from 1975. It would be great if an AI system could actually diagram it.
The diagram above is very bad however. I used to have my students do issue trees instead of essays and this one would get an F
But note that if each node has 3 nodes at the next level the tenth level has almost 60,000 nodes and the diagram about 100,000. The climate debate is easily that large so maybe AI is the only way to see it. People used to ask me if a computer could draw an issue tree of an active debate and I said no because computers cannot read. Now they can (accurately imitate reading).
i have never been able to get a funder because nobody wants to make the other side’s arguments clear.
But note that if each node has 3 nodes
=====
every question has possible answers: yes, no, and unknown. Everyone forgets that for all but the simplest problems the correct answer is unknown.
Climate change is not binary. Enthusiasts like to support outrageous catastrophe claims on the one hand, and dismiss lukewarm claims on the other. It is entirely reasonable to disagree on how bad climate change is going to be.
I wish that was well understood. But here in Wokeachusetts, I hear every day that we’re having an emergency. I look around – and just don’t see it. I must have a vision problem. Others see it- I don’t- despite having cataract surgery and my eye doctor says I now have 20-20 vision! I don’t get to Boston often- maybe there, they’re seeing the ocean boiling, I dunno.
Since the climate change they speak of is warming, discussing “how bad” it is going to be is illogical; a warmer climate is BETTER.
And there are many different kinds of unknown too.
Trying to download the file gets this error message.
Is there, if fact, a danger of the file being “corrupted or tampered with during download”?
Trying to open the file, once downloaded, seems futile.
Opened first time for me. No probs at all.
I’d appreciate if you could define “issue tree”. I did look at your article.
I believe that on multiple occasions I have argued online about climate with someone getting their answers from an AI algorithm.
The responses are suspicious, because they come up with lots of little facts that don’t really add up to a big fact–like a mind that doesn’t have genuine background knowledge. They can instantly come up with a list of citations you have to challenge, or else you give the appearance that they “fact-checked” you.
I got some indirect support for my suspicions when I came across a “Zooniverse” project, called ClimateVis, which is a solicitation for help from the climate alarmist public to process climate graphs and extract descriptions to be fed into AI algorithms. The link is https://www.zooniverse.org/projects/albasu/climateviz?utm_source=newsletter&utm_medium=email&utm_campaign=ISWCV12Aug2024
I would say that you will see, increasingly, these AI-based “fact checks” in your interactions on social media posts and YouTube comments.
The problem is, they are “Artifical”, but not particularly “intelligent.”
The do the English language well.. but that is just patterns and rules that can be programmed in.
To use the new Democrat go-to word in its proper application, this development is just
“weird”
adjectiveStrikingly odd or unusual, especially in an unsettling way; strange.
Can we expect AI to cut through the climate crisis/CO2 is causing everything BS that will dominate the data sources AI will use? AI digests only the data, commentary, and study results it is provided…
I am hoping AI might be the “truth seeker” that illustrates the science is not settled.
Any AI which concludes there is no climate crisis is discarded as defective, from what I have seen these guys only tolerate results which match their expectations.
I convinced the “Claude.ai” chatbot that the climate change scare was bogus giving it facts it couldn’t dispute. But that will disappear when I end the session.
for “science not settled”- I don’t bother with AI, I just read Koonin’s book
A good debater can argue both sides of any question without lie or falsehood. The notion that one can debunk a claim relies on small minds that cannot fathom infinity.
It is telling that disinformation is the illegitimate child of an unholy union between the CIA and KGB.
I had some early success with AI by pointing out that contradiction, not repetition, is the only reliable guide to the truth. The chatGPT engine then actually started to figure things out. However it then crashed and when service was restored it reverted to being a dumb lump. Believing the lie told often enough was true, regardless of contradiction.
This article is proof of all the bad things I have feared and suspected all along. AI is going to be nothing more than a propaganda machine, think 1984. They are taking something that could be beneficial and turning it into absolute crap.
It needs a lot of “training data” to establish a “consensus”. GIGO 🙂
At least in the chatbot: “Claude.ai” I could retrain it by giving it solar and climate facts it could check, and it recognized that that CO2 is probably not the cause of warming.
It will forget this when I end the session.
Large language models and neural nets are not AI as anyone would have understood the term “artifical intelligence”, until recently. They are more like artificial perception. They do not reason. They will never reason because that is not how they work.
So fret not, AI is not being turned into crap. A lot of people are being mis-sold neural nets, and it’ll take a while to flush through. But in the end we’ll all be a little wiser.
Can you give us examples of large language models, neural nets and AI to help us better understand what you are pointing out?
If John Cook is involved it will be pitched at a level to persuade 12-year-olds.
WTF is a “contrarian” anyway? At least one of the resident WUWT trolls loves to toss this one out, when he can’t back up his kooky pseudoscience.
“WTF is a “contrarian” anyway?“
A synonym for “REALIST”
The opposite of an “apologist” or “collaborator”.
Thank you! Like the marxo-democrats claiming they are “saving democracy”.
As they install a POTUS candidate who received not a single vote in the primaries. LMAO
They’re not going to stop with climate change. Next it’ll be the benefits of illegal immigrants, Islam and gender ideology. Before you know it, we’ll always have been at war with Eurasia.
oooo- you can’t say “illegal immigrants” in Wokeachusetts- you must politely refer to them as “migrants”
I will continue to call them by their correct name “Illegal aliens” They should all be deported forthwith.
It bugs me when some people say “we’re all descended from immigrants”. Sure, but my grandparents came here legally.
To be sure of the veracity of any model / program claiming to test bias then it must show it can tun both ways using test information which has a neutral domain as well as two extremities. There is nothing suggesting that this latest effort to square a circle is any better at it than any of the others ever tried since time began.
Why don’t these ‘experts’ put all the effort into proving that carbon dioxide is responsible for warming or even that the extra warmth today compared to the Holocene is of a different kind and therefore provably deadly.
What small minded people these academics seem to be but isn’t that always the case with consensus that steadfastly refuses to deal with proper science because none of the echoes have got the ability to do so and wouldn’t get paid anyway if they did attempt something original.
Interestingly, they seem to believe skepticism is as bad as “denial”. Don’t even think of questioning the experts!
“Misinformation about climate change poses a substantial threat to societal well-being.. “
Thine own lips have said it.
A substantial threat! Must be time for an inquisition!
“Automated detection of climate change misinformation offers a promising solution.”
Which would then have the erroneous messages instantly deleted? And, the name of the denier added to a data base?
“included a picture of prominent climate skeptics, including Anthony Watts, dressed as semi-naked Roman gladiators”
Anthony must have enjoyed that- I would.
Anthony’s words were “I haven’t looked this good since high school”. But I don’t think its something he wants to encourage.
I just got done with a paper by Nikolov/Zeller called Misrepresentation of Critical Satellite Data by the IPCC, which concludes that virtually all of the warming from 2000-2023 was due to solar forcing, and not by greenhouse gases.
https://tallbloke.wordpress.com/2024/07/26/nikolov-zeller-misrepresentation-of-critical-satellite-data-by-ipcc/
Here is what AI, which intercedes top of page at almost every search initiate, has to say about solar forcing. It leaves an Orwellian taste in my mouth.
Solar forcing is the hypothesis that variations in solar emissivity can bring about global climate changes. This hypothesis is undeniable in principle, since virtually all of the Earth’s heat comes eventually from the Sun. The problems come in finding episodes in the Earth’s climatic history where this has actually occurred.
Why did they use up so many letters to say “Re-education Camps”? Or even shorter still, “The Gulag”!!!