Have you ever scrolled through Twitter or Facebook and felt like something was... off? Like the posts are too generic, the comments too similar, the interactions too hollow? You're not alone. A growing number of people are subscribing to something called the "Dead Internet Theory." It's the idea that most of what we see online — from news articles to social media posts to YouTube comments — is actually generated by AI, not humans.
It sounds like a conspiracy theory. And maybe it is, partially. But after spending a week digging into the evidence, I'm not so sure it's entirely wrong. Let me walk you through what I found.
Where Did This Theory Come From?
The term "Dead Internet Theory" was first popularized around 2020 on forums like Agora Road's Macintosh Cafe. The core claim is that around 2016-2017, the internet shifted from a human-driven space to an AI-driven one. Bots, automated content farms, and synthetic media now make up the majority of online activity. Human interaction is the exception, not the rule.
At first, it was a fringe idea. But in 2026, with the explosion of generative AI — ChatGPT, Midjourney, Sora, and countless others — the theory has gained mainstream traction. A recent Pew Research study found that 43% of Americans now believe that "most content on social media is generated by AI." That's not a fringe view anymore.
The Evidence Is Everywhere
I started my investigation by looking at a typical Reddit thread. A post about a cute dog picture. It had 2,000 comments. I read through 100 of them. At least 30 were clearly bots — generic phrases like "aww so cute!" with no variation, accounts created in 2025 with no other activity. But what about the rest? Hard to tell. AI language models are good enough now that you can't reliably spot them. A 2025 study from Stanford showed that humans can only identify AI-generated text with about 52% accuracy — barely better than chance.
I then looked at news sites. I found a viral article about a celebrity scandal. The byline was a name I didn't recognize. I searched the author — no LinkedIn, no Twitter, no real presence. The article itself was generic, filled with clichés. I used an AI detector tool (which are also unreliable, by the way) and it gave a 78% probability of being AI-generated. That's not proof, but it's suspicious.
Why Would Someone Do This?
Money. The answer is almost always money. AI-generated content is cheap to produce at scale. You can generate thousands of articles, videos, or social media posts in a day. Even if each one makes only a fraction of a cent in ad revenue, the volume adds up. There are entire companies — many based overseas — that run automated content farms. They produce clickbait, fake reviews, and even entire YouTube channels. One investigation by The Markup in 2024 found that dozens of high-traffic websites were almost entirely AI-generated. The trend has only accelerated since then.
What Does This Mean for Us?
It's terrifying, honestly. If most of what we see online is fake, how do we trust anything? How do we know if a product review is real? How do we know if a news article is accurate? The Dead Internet Theory taps into a very real anxiety: that the internet is no longer a space for authentic human connection. It's a factory floor.
But I also think the theory overstates things. Real people are still online. I found genuine communities on Discord, niche forums, and in real-life meetups organized through Facebook. The key is to be skeptical, check sources, and look for human signals — typos, personal stories, specific details that an AI wouldn't generate. It's a skill we all need to learn.
I don't think the internet is fully "dead." But it's certainly sick. And if we don't start demanding authenticity from platforms and creators, it might only get worse. The next time you scroll, ask yourself: who wrote this? And do they actually exist? You might not like the answer.
