We had an assignment once to write about something we wished would happen before we die. Most everyone around me chose something cute like “win the lottery” or “travel the world” or some such thing. Mine was a bit more off-key. I wanted to see at least one conspiracy theory turn out to be true before the end. I wanted Paul McCartney, on his deathbed, to admit he’s really Billy Shears, and had been pretending to be Paul ever since the original died in a car crash back in ’66. I wanted Buzz Aldrin to admit the moon landing was filmed by Stanley Kubrick at an Air Force Base in Arizona. More than any of those, I really wanted the government to admit that Polybius was part of the MK-Ultra mind control experiments that were never truly closed down.
Finally, at last, I’m pretty sure I got my wish. Back in 2016, people in the darker corners of the web started whispering about something called the Dead Internet Theory. The claim? Most of what you see online in terms of posts, comments, and even arguments with your “friends” wasn’t written by humans at all, but by bots and AI. At the time, Ray Kurzweil was still promising us a utopia where AI would gently guide humanity into immortal robot bodies, and Hollywood kept warning us it would be more Terminator than Star Trek. I wasn’t buying either narrative. My cynical take was that if corporations like Google or Facebook got hold of AI first, they’d turn it into an ad-delivery system where the rich got richer, the poor got eaten alive, and the rest of us became Soylent Green with Wi-Fi. Turns out, even I was too optimistic.
Fast-forward to a few days ago. My Threads feed, as it often does, shows me a girl who posts a picture of herself in an ill-fitting tank top, captioned, “I’m lonely, won’t somebody play with me?” Okay, fine, except that this account has like six posts, a bio claiming to be an “influencer” without links to any other type of content, yet as somehow in this time garnered 1.5 million followers. The picture alone had nearly five thousand comments in nine hours. Out of curiosity (okay, petty rage), I dug deeper. Turns out, the 20 or 30 most recent comments were people wondering how someone with zero other social media presence and only six posts had amassed over a million followers. Then came the kicker: copy-paste “I’m lonely, won’t somebody come play with me?” into the search bar, and … boom … she has about 200 “twin sisters,” all equally lonely, equally over-followed, and equally suspicious. If this sounds like the setup to a Black Mirror episode, congratulations, you’re paying attention.
I mean, okay, we know there are a lot of bots out there. My original draft of The Halferne Incubus dates back to 2002 and was inspired by an incident where I found myself in an IRC chat room with 30 people, 29 of whom were bots with various nefarious purposes. 1.5 million of them working together and replicating themselves, though? This almost goes beyond the dreams of a shlub, part-time science fiction author, and headlong into Dead Internet Theory. It seemed “out there” at the time, but now makes perfect sense.
It’s not just a gut feeling. According to Imperva, over 52% of internet traffic in 2016 was bot-generated. By 2022, that number had climbed to nearly 50% again after peaking even higher. Imagine, every time you click, it’s basically a coin flip whether you’re interacting with a human or an automation script. And that’s just what corporations admit to. If they’re saying 50%, you can bet the real number is closer to 70–80%. Corporations don’t lie, they just tell the truth in fractions that won’t tank their stock prices.
What I’m getting at is, human voices were smothered under a rising tide of bots, corporate astroturf, and AI-generated filler content. What you’re scrolling through on Twitter every morning during your “personal time” is the content equivalent of a house of cards assembled by automation. Now, I know the contrarians among you are thinking, “Pfft, that’s silly! The girls all have photos, and they respond to my comments, of course they’re real. They just want your money.” Yes. Exactly. That’s what they want you to think. Given the amount of my datapoints that must be floating around out there, our silicon overlords could create perfect AI girlfriends for me at the rate of thousands per minute now, after all. Wait … let me ask ChatGPT for one …
Need more proof? Do you have a minute to talk about “Shrimp Jesus,” our shellfish savior? In case you slept in that day, back in 2023, social media got flooded with thousands of accounts posting pictures of AI-generated images of Jesus as a shrimp, and each post seemed to immediately racked up thousands of likes, shares, and “Amens.” No, there is not a prawn-prophet? Instead, researchers at the University of New South Wales called it out as AI-driven content farms engineered to find and exploit platform algorithms. All of it was contrived. None of it was human-driven. It was just a few thousand AI bots cranking out memes and simulating engagement until *DING* they gamed the system and went viral as a test to find the gullible among us who wanted cocktail sauce with our communion.
The Financial Times recently dubbed our ecosystem the “hostile Internet,” a landscape dominated by spam, scams, and AI-generated garbage. In other words, you never know if the person agreeing with you online is your neighbor or a server farm in Singapore. This isn’t paranoia. It’s a business model. Ad revenue needs eyeballs. Eyeballs need content. If humans can’t produce content fast enough to make the platform profitable, then cue AI to fill in the gap. It doesn’t matter if the content is true, useful, or sane. All that matters is that you keep scrolling like a hamster running on a dopamine wheel.
So yes, I believe the Internet is dead, but like any good zombie, it’s still wandering around, gnawing on brains and mumbling “Like, share, subscribe.” We’re basically watching The Muppets, only everyone’s Kermit, nobody’s Jim Henson, and Statler and Waldorf are just AI sock puppets telling you to buy a hot new secret crypto stock. Personally, I’m relieved. I thought the Internet was full of complete morons with the intelligence of warm pudding. Turns out, that’s just what makes me scroll further looking for something to engage with.
We no longer live in a media-saturated world. We live in a media-replaced world. Bots don’t need to perfectly imitate us, they just need to be “human enough” to pass the sniff test. If you find yourself nodding along to a Twitter bot’s hot take…well, congratulations, you just became useful to them.
Still think I’m overreacting? How do you know this post isn’t itself AI-generated? Maybe my typos are red herrings to make me seem human. Maybe they’re a secret code from the underground resistance. Maybe I’m here to convince you of the truth. Or maybe I’m here to tweak your contrarian flat-Earth nature and convince you it’s all nonsense so you spend another ten minutes doomscrolling and hitting some faceless ad network’s engagement goals.
Go ahead. Try to prove I’m human. I’ll wait. Either way, you just added another data point to the algorithm.
Update: Sam Altman, probably after reading this blog post, has announced a reversal of opinion and is now “Suddenly Worried Dead Internet Theory is Real.” Which means, technically, I’m thinking ahead of the CEO of an AI company, and we’re both behind the scam artists.
