Now Reading: SLOP II – Pure Slop History
-
01
SLOP II – Pure Slop History
SLOP II – Pure Slop History
An impressively lazy impression of history
Paladins: Learn History is an app that is done with you as soon as you have paid. While it’s ostensibly a history learning app styled after Duolingo, the promise of learning history is actually just bait for the app’s hook – which is using the conventions of app design to get inattentive users to tap the 7-day free trial button, but it’s a 7 day trial… for a $100 /year subscription.
Like the pin-sized bit of metal which ends a fish’s life, the hook is frustratingly simple – and you can almost smell the devs gulping and rubbing their hands waiting for some poor sod to misclick that hundred buck button.
Once they’ve got that purchase, or the lesser prize of a (still pretty wild) $15 per month fee, the actual product is the minimum needed to avoid being legally a scam – but it is a lower minimum than I realized you could even go! Everything in the app – the images, the maps, the voiceover, the text itself, the quiz questions, even the advertising surrounding the app – is blatantly generated by AI software. There is nothing but the basic code behind the app that I can believe was not made with AI – and even then I got my doubts. Every effort has gone into minimizing effort and minimizing cost.
But the thing which this cheapest-possible-product claims to be is History, the memory of the past of our world, an incredibly important thing (especially to nerds like us at the VALUE foundation!) For the fishermen, the bait is a means to an end, but I’m interested in the worm what hungers the fish.
🎣
For some basic context, it’s published by Deepstash Inc, a grindset-culture publisher who mainly traffic in AI-generated summaries of self-help books to “BECOME THE SMARTEST IN THE ROOM.” Based on reviews from the app’s android store and app store pages, the game has actual users, who complain of a shortage of content – which means there are people who have consumed all the history the app delivers, and are frustrated there isn’t more. It’s got about 1.3k 5 star reviews on Google Play, and looking at them, it’s not all bots, which means people are actually using this app. Its telling of history is becoming part of the public historical memory – it has an impact.
It’s actively advertised on Tiktok and the Facebook ecosystem with AI-generated video shorts, which place it clearly in its larger genre: online slop history. Linked are two ads for Paladins, and two slop history shorts for comparison.
The audience interested in the latter are clearly the target audience of this app, and that’s a big market! Popular channels in the slop history genre can see viewership in the range of millions, and at the extremes tend to dip into the pseudohistory classics of biblical mythology or ancient supertech that have always been big across the web, though Paladins is by far on the more restrained side.
So, what’s actually inside? You boot up, and you complete a short personality quiz to assign you your ‘Paladin,’ an avatar from one of the app’s historical characters. After this initial bit of gameplay wets your whistle, you’re taken to an impassable subscription panel. If there’s a way around, I haven’t found it, and here the app tries its darnedest to get a user to agree to a 1 year subscription by pressing the 7 day trial button. I go with the 15 per month subscription.
[Author’s note: During the course of writing this article, Deepstash has updated the UI and added a little skip button to the account creation panel. Creating an account still forces you to subscribe.]
After that, you pick a historical character to follow, which leads the user through a linear biography of that figure. These are, in every case, images of a very badly made map or bizarre AI generated image, accompanied by automated text narration giving a cheerful and basic summary of these characters’ lives, with more analogies than a box full of analogies.
Every other page has a ‘deepdive’ button, which opens a new section of text with sometimes more information, sometimes the same information you just heard. These are broken up by unfailable binary or 4-option choices, where the game asks ‘what do you think will happen next?’ Being the rebel that I am, I like to pick the wrong answer. 😏
At the end, you’re alerted by a spinning orange starburst that it is:
You get a very short and very easy quiz on what was just covered (And sometimes information which was not in the chapter!), which consists of matching character with a statement, or choosing from another binary or 4-option choice, with no penalty for failure if you’re on pro. If you’re not paying, you watch an ad. Each story is this, repeated 4-5 times, followed by a ‘battle’, which is a timed version of the same quiz.
Finish that and you unlock that ‘Paladin,’ which just gives a score bonus to your next quiz battle and another avatar. Congratulations, that is the entire game which you may have paid 100 dollars to play for a year! It’s no Baldur’s Gate, I’ll tell you that!
In excessive fairness, Paladins is marketed as a once-per-day self-improvement app, to fit in a bathroom break. Its ads strangely frame it as a way to get people to like you, or to trick people into thinking you’re a genius.
The whole thing is upfront about aping mega-popular language app Duolingo, it uses a Duolingo style bubble-interface, it has the chimes, the progress bar, and Duolingo’s tinny fanfares. How it teaches history is actually very similar to old educational approaches you may have encountered in school – multiple choice quizzes, short linear narratives centered on a famous individual, and a description of ‘what happened in the past’ as statements of certainty. All told, it’s just a bad game, and that defeats the whole purpose of gamifying learning – but even then, its history is very odd.
Note the similarity in fonts, progress indicators, and rounded-square interface elements.
Before we get into the more serious part, let’s have a few laughs:
Look at Pompey! Look at his face! Why did they do him so dirty?!
And the maps, how did this even happen?
Sometimes the text references images which don’t actually appear, but must have had an equivalent in the training data, like the ‘red dots’ or ‘angry arrows’ referenced here:
Maybe the most visibly absurd, the way male bodies are presented in the app is hilarious. Nearly every man of note becomes a chiseled male model, and if they were a tough guy, they have truly absurd musculature that should never be constrained by shirts!
It’s funny, but it points to one of the more concerning parts of Paladins‘ historical perspective. The app has a deep love affair with hypermasculine, physically powerful authorities – their Leonidas chapter features framings of the events of the Second Greco-Persian war that border on pro-Spartan apologeia, glorifying the Spartans’ discipline and how they ‘saved Western Civilization.’
The whole section is full of loving descriptions of the power, honour, and masculine might of the Spartans, while framing their system of mass slavery as a tradition they were heroes for defending. This, combined with the centering of all narratives on a handful of great men (in the historical sense) and ads like this one give a concerningly fascist tone to their narratives.
Is this Paladins‘ stance? Or is this an unfortunate coincidence of their AI use, serving up existing messages? I don’t know, because this is also a really popular framing for pop history – mighty warriors of action and great men of history and buxom babes. Fascistic strength-worship is the framing of the most popular depiction of Leonidas’ life, Zac Snyder’s 300. Yet, that I cannot say whether this is intentional or unintentional authoritarianism can speak only to irresponsible history-writing.
Going from the large scale problems, to the smaller scale, we can look at Paladins‘ narration of Rosa Parks’ life.
Rosa Parks
For that, let’s look at Rosa Parks’ appearance in Paladins. Her whole story is told in a strange mix of ‘this isn’t the story you know,’ and very typical tropes of a sanitized telling of the Civil Rights Movement, which combined with the surreal cartoons of AI images already puts a bad taste in my mouth. But part 2 of her storyline starts with an anecdote of a little white boy pushing her to the ground as a child, and credits it as when she ‘first learned that dignity was something to stand up for.’ This, and her grandfather (a formerly enslaved man) defending her family from the KKK, are framed as what made her politically conscious.
But the former event plays a pretty small role in her own account of her life. Within Mrs. Parks’ autobiography, Rosa Parks: My Story(1999), this event is relayed because it was what caused her guardians to move her to another school away from the white neighborhood, but it had no significant role in her political awakening (pg. 48). The events are the same, but Paladins’ framing has significantly modified its meaning.
Yet she gives a lot of importance to a different, similar event – when a white boy attacked her brother and she threatened the boy with a brick. What that event meant to her was not an easy aphorism like “learning dignity was worth standing up for”- she remembers how much it hurt when she felt her grandmother took the white boy’s side over hers, when she warned her that she could be lynched acting like that, and her frustration that she was being scolded for doing the right thing (pg. 22-23).
These are not hard details to find, this anecdote opens the second chapter of her autobiography, which is accessible for free on archive.org and is short enough that it would take a human writer less than a day to read the whole thing. This is one of a number of strange re-framings the app’s history creates in service of a certain inspirational tone. But AI generation means nobody on the development team so much as looked at her autobiography – and an ordinary user who wasn’t scrutinizing the app would have no way to know the misrepresentation.
It is important to remember, Rosa Parks was a real person, and to tell her story or anyone’s story is a responsibility. The most basic level of historical responsibility is trying to get the facts right even when you may make errors, but for the sake of profit, the creators of Paladins shirk that responsibility. Instead her life is rendered by ChatGPT, by mathematical average, and no one who could check its work is employed, because that’s too expensive.
I was left asking after playing through multiple of these storylines: Have I actually learned? The history was so clearly generated by software that is well known for hallucinating data, and I can recognize some of the errors, so I’m certain there’s far more that I don’t notice when the topic is one I’m unfamiliar with. How can I take information from Paladins with me, if I can’t trust it? By using Generative AI in the place of a human author, Paladins destroys the users’ trust – and according to their terms of service, Paladins are not legally liable to provide accurate or factual information:
As it stands, this app undermines any potential for education because a user cannot reasonably trust its facts. That’s bad education.
Can GenAI write history?
But even if the generated text was more prone to accurate facts, if the tone was more appropriate, if the images it generated looked more like the real people, Paladins still couldn’t escape the rot at its core. That rot is using Generative AI to write history, because due to the way the tech works, it is only capable of generating things which look like history, but it cannot do history. That’s a bit confusing, let me explain.
GenAI’s like ChatGPT are built by scanning vast amounts of text and images to create a mathematical model of likely combinations of pixels, words, and letters. This can be used to generate new text or images which resemble to a very impressive degree its training data, and the software can even reproduce specific different tones in this way. Using this, it can be set up to respond to prompts in a conversational format, which is how users can seem to ‘have a conversation’ with the AI. It is a super impressive mimicry machine, and just like any good mimicry, it is hard to distinguish from the real thing. It’s to the benefit of AI companies to hide the nuts and bolts of how it works, because when you see it in action, it really does seem like you’re talking to a thinking entity.
At its core though, this isn’t history. The act of doing history requires thought, because even at its most basic, history is the process of attempting to construct a narrative about the past based on the evidence it left behind – documents, eyewitness accounts, physical artifacts. This can even be making popular history, like a historical film (like Schindler’s List) or a stylized autobiography (like Maus). These are always flawed representations of the real past, but that is a fundamental limitation we gotta work with. This is a task that there is currently no software which can actually automate, but the products of this task – works of history-writing, can be imitated convincingly. That imitation is a representation of a representation of the past, and that becomes the hazy territory of what is called ‘simulacra’.
The term’s most famous in its use by the philosopher Jean Baudrillard, and basically means a representation of a representation of something real, where that more distant image’s relationship to the real thing has been lost. Imagine I find a fossil in my backyard and say: ‘wow, it’s a bone! I wonder what sorta creature had this bone?’, and so I try to rebuild the dinosaur from one of its bones.
Very reasonable, but I’m mistaken, fossils resemble bones but they’re actually minerals filling in where a bone used to be – they’re basically a cast outline of a bone. If I don’t realize that this is an image of the bone, and treat it like a real bone, my attempt to reconstruct the real dinosaur will be a creature that never existed: A dinosaur whose bones are made out of rock. This is what Baudrillard calls ‘hyperreal’, the ‘real thing’ we try to reconstruct from a simulacra.
Because the text in Paladins looks like it was actual history-writing, you can come away thinking this is a representation of the real past, but the software making it has no conception of the past, just what looks right.
Returning to the original point of making this app, I have no reason to believe the creators of Paladins are actually trying to teach people history – it’s bait and a hook. While worms are a popular traditional form of bait, you don’t need real food to catch fish, you can use a lure that looks like something the fish wants, and the only person unhappy will be the fish – and anybody who cares about fish getting real food.
So, Paladins is one drop in the big bucket of ‘slop history’, and the reason it’s a problem is pretty clear now.
Slop history is:
- Made for generating exploitative profit
- History made without concern for its accuracy
- History which reproduces existing problematic tendencies
And with the advent of Generative AI:
- ‘History’ which isn’t made by people thinking about the past, but made by advanced mimicry software which makes things that ‘look plausibly like’ history.
This is a genre that we all will have to reckon with. If we accept as ‘history’ stuff that does not attempt to represent the past, but just looks about right, our society becomes less understanding of our real history – it all trends towards being just another type of fiction.
Now, does it matter whether we’re remembering the real past or a sellable, cheap version? Does it matter if Mrs. Parks’ radical life is misrepresented? Does it matter if our society uses AI as a substitute for thinking? I sure think it does, and I think the people who benefit from everybody else not having a tangible grasp of the past or access to real, good history are not people I want to see winning. I certainly don’t think we should be accepting all that, for the sake of a bit of profit in the pockets of a few scammers – and that is what slop history is.
Next time, we’ll look at the future of ‘slop’ the concept, as it moves through linguistic tides. Until then, let melting Pompey see you off!
Hello, I’m Jacqueline Saunders. I’m a student here at the Research Masters Art Literature and Media, but I’ve previously studied a bachelor’s in History, and I’m particularly looking into the intersection of digital culture and historical imagination. In my free time, you can find me hanging out with geese, writing, or gobbling up frankly unreasonable volumes of history and philosophy. I have a fondness for striking fashion, how invisible concepts shape our lives, or games about architecture.
If you want to get in touch, you can reach me at: linathsau@gmail.com
Related Posts
Stay Informed With the Latest & Most Important News
Previous Post
Next Post
