Dead Internet Theory: Is the Internet Already Dead?

There's a theory that the internet died sometime around 2016. Not literally. The servers are still running. But the version of the internet that felt alive, chaotic, and genuinely human? Gone. Replaced by something that looks like the real thing but isn't.

It's called the Dead Internet Theory, and it started in a post on a fringe forum called Agora Road in 2021, written by someone called IlluminatiPirate. The post argued that most of the content we see online is no longer created by real people. It's generated by bots, AI systems, and corporate content farms, all designed to simulate the feeling of a living internet without any of the actual humans behind it. A "hot air balloon with nothing inside," as the original post put it, compared to the internet of 2007 and earlier.

The Atlantic covered it that same year in a piece by Kaitlyn Tiffany titled "Maybe You Missed It, but the Internet 'Died' Five Years Ago". Her conclusion: the theory is wrong, but it feels true. That tension is worth sitting with.

Why It Feels True

Because the numbers back up the feeling, even if the conspiracy doesn't.

Imperva's 2025 Bad Bot Report confirmed that for the first time in a decade, automated traffic surpassed human traffic. Bots accounted for 51% of all web traffic in 2024. Bad bots specifically, the kind designed to deceive, scrape, and manipulate, made up 37% on their own.

A pre-print study from AWS found that 57.1% of all sentences on the web are machine-generated. Nearly 1 in 7 websites hosts AI-generated content. And that was before the generative AI explosion fully hit.

The internet we grew up on had texture. Personal blogs. Arguments that went nowhere. Photos that were slightly blurry because someone just took them on a real phone in a real moment. That texture is mostly gone now. Replaced by content that's optimised, generated, and endlessly recycled.

It's Not a Conspiracy. It's a Business Model.

The Dead Internet Theory tends to get lumped in with tinfoil-hat territory because its wilder versions involve government psyops and coordinated manipulation. But you don't need any of that to explain what's happening. The boring explanation is bad enough.

Platforms rewarded engagement above everything else. Bots drive engagement. AI content is cheaper than human content. So the incentive structure quietly filled the internet with non-human noise, and nobody in a position to stop it had a financial reason to do so.

Research published in early 2025 on arXiv put it plainly: social media platforms increasingly use algorithms and machine learning to curate content, drive engagement, and maximise advertising revenue. While these tools enhance scalability, they prioritise virality and consumption over authentic communication, contributing to the erosion of trust.

That's not a conspiracy. That's just capitalism doing what capitalism does when nobody's watching the door.

So What Do We Do?

The theory, in its most literal form, is probably wrong. Real humans are still out there. They're still posting, still connecting, still having genuine conversations. But they're increasingly outnumbered, and the platforms built to host them have stopped caring about the difference.

The only real answer is to build somewhere that cares. A space where every account is a real person, verified not once at signup but continuously. Where the friction isn't a bug in the design, it's the whole point.

The internet doesn't have to be dead. But keeping it alive takes effort. That's what Hi Friction is for.