The Internet Is Not Dead, But It Is Going Underground
On the "Dead Internet Theory", conservative astroturfing, and AI.
A dangerous conspiracy theory is making a popular resurgence, including and especially among left-wing commentators.
If you’re online as frequently as I am (first of all, get help, but also), you’ve probably seen an alarming amount of AI-generated content in your feeds. As DALL-E, Midjourney, and other image-generation tools become more accessible and easy to use, more and more of our digital landscape is being populated with uncanny images like the ones documented by my TikTok mutual Hazel below. Worse, most of social media seems to only exist for content farming by bots…
Tiktok failed to load.Enable 3rd party cookies or use another browser
Generative AI has been the topic du jour for the past year, with no signs of leaving our collective consciousness. A few weeks ago, I wrote about the ways in which the recent AI proliferation has compounded with our lack of media literacy to the detriment of democracy (aren’t you glad you subscribed??) This week, I want to talk about the Dead Internet Theory, what it gets right about the state of our culture, and what it gets wrong.
Interest in the “Dead Internet Theory” has spiked over the past few months, but what even is it? In short, it’s the belief that the Internet is now mostly, or entirely, populated by bots, and that real free speech by humans on the Internet has been eroded by corporate domination, algorithmic echo chambers, and government censorship. The evidence? Seemingly, most of the comments you see under viral posts are bots. And lately, we’re seeing what appears to be bots creating content, which is then commented on and shared by bots, with little human involvement. The AI is eating itself, and we are all doomed. I’ll get into a bit later why I think this is not a good analysis of the modern Internet, but first, I’ll concede there is some credibility to the idea that bots have a weird amount of influence over our online experience, especially in our political discourse.
In a new podcast series, Tortoise reporter Alexi Mostrous and producer Xavier Greenwood take a deep dive into the online hate-farming that took place during the Depp v Heard trial. I have long held support for Amber Heard and, while the trial was ongoing, I was honestly shocked to see just how much content was being produced in favor of a man who, allegedly, objectively abused her. “Surely,” I thought, “there’s no way this many people actually believe this, right? I know that there are lots of misogynists in the world, but people I know and trust are making disgusting content in favor of an abuser!! How do people not see what’s really happening?” The trial proved that, yes, many people need to take a Feminism 101 class, but as the Tortoise reporting lays out, that hate didn’t always come from humans. Rather, a good portion of the anti-Heard hate came from bot accounts based in Saudi Arabia.
It’s not necessarily that all of the Heard hate came from bots. But if you know anything about how content algorithms work, you know how easily it is for movements to be astroturfed. A few dedicated accounts can get a topic “trending”, which attracts some bots (who are programmed to latch onto trending topics and SEO-friendly keywords), then a few more humans hop in because “hey a lot of people seem to be talking about this, what’s the deal?”, so more bots jump in with them, and the topic keeps cascading into popularity. As the content trend grows, it creates a culture of hatred with plenty of incentives to buy into it; “lots of other people are hating on Amber, so it must be okay!” (people made good money on pro-Depp content, too).
It’s alarming just how much of our beliefs are based on an extremely diluted sense of what’s really going on in the world, how many Internet “users” are actually bots, and how much human behavior can be controlled by a few bad actors with bots on their side. This lack of agency over what you believe and why is scary to confront, so most people bounce off, rightfully clinging to the idea that they’re a rational actor, not a victim of the floor of human behavior being tilted beneath them. However, others lean into these ideas, and eventually come to the conclusion that actually, most of the Internet is “fake”.
It’s a deliciously tempting idea: that all, or at least most, of the users you see online who disagree with you are simply not real. But is this a new idea? Where have we seen this kind of “digital astroturfing” before?
…Well, lots of places. Numerous studies have shown the ways in which (for example) Facebook pages pivot their content from innocent memes to dangerous ideologies. On a smaller scale, you have pages devoted to spiritualism pivoting to anti-vaccine conspiracies. On a larger scale, you have entire operations where Facebook pages devoted to sports, memes, and funny videos suddenly pivoted to extreme anti-SJW content in times when that content was popular, only to go right back to sports talk a few weeks later. The logic is that you get on people’s feeds by reposting viral content from elsewhere online, getting people to follow you for more silly memes, so that when you pivot to politics, you end up on their page without anyone engaging with a single political post to begin with. Facebook acknowledges this problem (dubbing it “abusive audience building” or “inauthentic behavior”) and occasionally takes steps to remove these accounts. But with more actors engaging in this sort of activity, it’s becoming increasingly difficult (maybe even impossible) to stop.
Many of these operations take place outside of America—talk to any brown person you know and they’ll tell you how their auntie was radicalized by a WhatsApp meme group chat—but the stories that typically make American headlines involve specifically a foreign actor influencing our politics. Ideas like “bots from Russia/China/the Middle East/etc. are influencing our politics!!” map on well to Western xenophobic ideas of invasion or “Great Replacement”, and I don’t want to do that here. Instead, I’ll point your attention in the direction of numerous domestic efforts to sew misinformation and hate in favor of far-right ideology, including…
Cambridge Analytica: that time Facebook sold psychological profiles of millions of Facebook users to a data firm who was then paid by the Trump campaign and Brexit “Leave” campaign to show those users ads featuring far-right content
The Epoch Times: that time a far-right newspaper created hundreds of fake pro-Trump accounts to spread their ideology across multiple languages
The National Federation of Independent Business (NFIB): that time an organization claiming to represent workers supported rolling back child labor protections
Sinclair Broadcast Group: the alarming reality that many local news stations are beholden to a conglomerate that pushes conservative programming
If there are any cases of leftists using this strategy to win over conservatives, I haven’t heard of any, although I’m more than willing to be proven wrong on this! In general, though, wealthy conservatives are the ones who a) want to maintain the current status quo by pitting the working class against each other and b) have the economic means to make sure that happens using all the various forms of media they can. But that’s just my own Marxist analysis; feel free to disagree with me in the comments!
Clearly, AI isn’t the cause of this astroturfing phenomena per se, it’s only allowed for easier access to this political strategy. One no longer needs to spend hours scraping content for your pivotable meme page; one can just create Midjourney images using SEO-friendly keywords. One no longer has to create and maintain a few dozen social media profiles manually; a computer can now make you thousands fake profile photos, bios, and even text posts. Look, even I can do it!
While this phenomena of “bad actors creating thousands of bots on the Internet to sway opinion” seems to be accelerating as time goes on and AI becomes more advanced, most of the people who discuss the Dead Internet fail to mention the inherent right-lean in the actual cases of it, several of which I mentioned above. I bring this up to shine light on yet another absurdity, one that nobody ever seems to talk about when it comes to this theory: the fact that the “Dead Internet Theory” was actually created by a far-right conservative.
The “Dead Internet Theory” was popularized in a post on the Agora Road forums in 2021. The text is full to the brim with 4chan-speak; the poster describes themself as an “oldfag” (a person who had been using 4chan for a long time), as being into “loli and feet”, and “among the first right wingers who were such before it was cool” (imagine being so conservative that you see yourself as morally superior to 4chan users, but in an even-further-to-the-right way). The post lays out some tenuous connections between some AI companies and government agencies like the CIA (which, yeah, of course the government is looking into AI to some level, it being a national security threat and all), but among the first pieces of “evidence” that they bring to the table are a) some people they used to talk to on the Internet don’t post there anymore, b) memes like “Raptor Jesus” and Pepe the frog look kind of similar, and c) “fiction has become sterile”, a complaint so generic every boomer has said it at least once. It’s some classic Internet brain-rot stuff.
Most prescient to me was their notion that “pedo activism” (their term for the LGBTQ+ movement) has only gained popularity in recent years because it serves as popular “fetish content” for 4chan users…
"Why does the real world bend over backwards to accommodate our weirdest fetishes? It's as if everything is going "Look, look! I created this for you! I made it real!" in an effort to keep us within this world. The results of this are devastating to society, to people, to civilization. Simply put, trannies are a thing because Anon fapped to doujins of cute boys in dresses. Once it was an impossible fantasy, not to be taken too seriously. Now it's grim reality. Again: it's as if the real world is using imageboards as a template on what to be and what to do." -Author of the original Dead Internet Theory forum post on Agora Road
While some of this language might require a visit to Urban Dictionary for digital non-natives, it shouldn’t take a genius to figure out that this person (or multiple people, it’s not entirely clear how many contributors there were to the post) is an ultra-conservative who thinks that LGBTQ+ people are unnatural, dangerous to society, and largely based on pedophilic fetishes. The claim that “transgender people are only gaining rights because the CIA saw that 4chan users were jerking off to trans porn and so they astroturfed a whole civil rights movement in response” should really be enough to dismiss the Dead Internet Theory entirely. I think that anybody who brings up the Theory absent of this context is committing journalistic malpractice.
It may seem like I’m doing the thing that all terminally-online leftists do; saying that because someone is wrong in one area (e.g., they are transphobic) then they must be wrong about all other things. I try to avoid that kind of petty finger-pointing, as I hope is clear in my other writing. I’m mentioning the theory’s authors’ political affiliations not to discredit them solely on the basis of being conservative, but to point out how the Dead Internet Theory is inextricable from the kind of narcissism, individualism, and aversion to new ideas that’s characteristic of the far right.
Here’s a phased progression of increasingly-extreme ideas that one could come to if one buys into the Dead Internet Theory:
“A lot of the people posting about social justice movement X appear to be fake accounts; this must be because bots latch on to trending, SEO-friendly key words.”
“A lot of the people posting about social justice movement X appear to be fake accounts; this must be because fewer real people actually believe in the cause than what it seems like online.”
“Social justice movement X must be entirely astroturfed, because nobody in their right mind would believe in this cause.”
“The government (or shadowy group of your choice) is manipulating the public en masse to promote social justice movement X, whose real goal is to destroy society.”
“Most of the Internet is bots, so it must be that everybody who agrees with me on X is a real human, while most of the people who disagree with me on X are not real people.”
“When I’m on the Internet, I should behave as though I am the only real person I know, and literally everybody else is a robot.”
Notice how even though the top observation is pretty plausible but the last seems insane, the progression of each idea to the next isn’t that extreme of a jump. I shouldn’t need to explain why this level of solipsism is poisonous to political discourse, but this is apparently where we’re at, so here I go anyway.
“Everyone who agrees with me is a real human, while everyone who disagrees with me must be a bot” is an incredibly dangerous worldview to have; it dehumanizes people, allowing us to, at best, ignore the potentially-valid political concerns of our political rivals, and at worst, actively harm or kill them. And yet it’s infected political discourse on all sides. On the right, the term “NPC” (literally “non-playable character, the soulless dialogue script-spouting stack of polygons) is wielded against those seen as even vaguely progressive. On the left, we regularly imply that simply by looking at someone’s Twitter bio, we know not only what a conservative believes, but why they believe it (they’ve been corrupted by the systems of racism, patriarchy, et al…and if it’s a Black or female conservative, then it’s simply internalized racism, internalized misogyny, et al.)
This worldview is absolutely poisonous to our ability to maintain real relationships and create real change in the world. In many ways, terminally-online leftists talk about “the revolution” in the same way Christian fundamentalists talk about “the rapture”; a complete shift in the power balance of society coinciding with the culling of the (ideologically) impure.
I do maintain that conservatives tend to be more actively violent, however. They shoot Bud Light cans with guns because they can’t yet do that to actual trans people. They threaten their political opponents with murder. They call us NPCs, and remember, what do gamers do to NPCs in video games? They take advantage of them. They steal from them. They run them over with our cars in GTA. They shoot them in the head to see if the game’s programmers accounted for this radical action, or just to see some funny ragdoll physics. Worse, it’s typically conservatives who claim that the Left is secretly in control of the media and inflating its numbers in the way that it’s very clear that the far-right is doing. As usual, every accusation is a confession!
I don’t believe in conspiracy theories. Most of the “strange” phenomena in society is entirely explainable via the incentive structures of capitalism, colonialism, some kind of bigotry, or just plain human stupidity. I have no need to chalk things up to the the inscrutable actions of shadowy figures or lizard people (which by the way, are both antisemitic tropes), because I have intersectional feminist analysis on my side!
Yes, the Internet is made up of a lot of bots now. Yes, we need to be ever more vigilant about the ways we might be manipulated by the media we consume. But that does not mean the Internet is dead. If anything, it’s more alive than ever; you just have to look for it.
I’m excited about the return of personal websites. Long before social media, we learned HTML and created our own weird little personal webpages, linked to one another by webrings, and it was fun. Getting all your content fed to you through the medium of Facebook or Twitter of TikTok is not fun, and given the incentives for Facebook to increase profit rather than increase democracy, it’s definitely not pro-human. Stumbling on a neat little webpage can be very fun, though. CJ the X just posted a fantastic manifesto on the return to Web 1.5, which I found incredibly enlightening! Yes, the web going “underground” makes certain things harder to find, but it makes things so much better than a) platforms controlling everything and subsequently b) bots being able to take them over.
I myself have a personal website, readers! It’s fairly primitive, but it’s the hub for all the various things I do online. Go check it out if you have a moment :) www.ThatAnnaMarie.com
Fare well and stay vigilant!
Currently Reading
Parker Malloy’s opinion piece on the Taylor Lorenz/Chaya Raichik interview.
Watch History
An insanely impressive Tirrrb video about the Spiderverse movies!
A fascinating peek into conservative culture by a trans woman who managed to sneak into CPAC.
A deep dive into the far right’s infiltration of pro-Palestine content, which I also touched on in a video a few weeks ago.
An incredible look into James Baldwin’s takes on gender and sexuality.
And now, your weekly Koko.
That’s all for now! See you next week with more sweet, sweet content.
In solidarity,
-Anna