Is the ‘dead internet’ theory coming true? New Stanford research calculates exactly how far we are—and it’s alarming

It’s official: the robots are taking over. Taking over the internet, that is.

Conspiracy theorists have long discussed the “dead internet” theory, which reasons that online spaces, once entirely populated and filled with content created by humans, have slowly become dominated by bots posing as people. The more extreme conspiracists allege that this transformation is deliberate, with governments and corporations using the bots to manipulate public perception.

With the rise of AI since ChatGPT’s debut in 2022, the dead internet theory—or at least some version of it—has sounded more and more plausible. Now, according to a recent study, it’s closer to coming true.

The study, a collaboration between researchers at Stanford University, Imperial College London, and the Internet Archive, sought to find how much text on the internet is AI-generated. To do so, they used the Internet Archive’s Wayback Machine to compare web pages published between 2022 and 2025, using several prominent AI-detection methods to see just how much online content was created partially or entirely by AI.

Their results show that as of May 2025, more than a third of all new websites were AI-generated or AI-assisted at 35.3%. That includes 17.6% of all newly published websites being entirely generated by AI.

The study’s findings corroborate other data, including Cloudflare’s report that nearly a third of all internet traffic over the past year comes from bots and Imperva’s claim that 2024 saw automated traffic surpass human traffic for the first time.

“I find the sheer speed of the AI takeover of the web quite staggering,” Jonáš Doležal, one of the study’s researchers, told 404 Media. “After decades of humans shaping it, a significant portion of the internet has become defined by AI in just three years. We’re witnessing, in my opinion, a major transformation of the digital landscape in a fraction of the time it took to build in the first place.”

Defying AI stereotypes

While AI-generated websites are becoming more and more common, they may not be as harmful to the internet as critics fear. The study also tested AI-generated content’s effect on the internet via six hypotheses centered on common critiques of AI-generated text, only two of which were confirmed by the research.

The study confirmed that AI is contributing to semantic contraction, or the reduction of diverse viewpoints online, and to a positivity shift, through which online writing is overall becoming more sanitized and artificially cheerful.

But so far, AI is seemingly avoiding other suspected negative effects. The study didn’t see evidence of increases in rambling text with little to no substance, a single generic writing style, a lack of cited sources, or—perhaps most shockingly—the spread of misinformation among AI-generated content online.

The study’s authors told 404 Media that they’re working to turn their research into a continuous tool, so internet users can stay up to date on just how dead that “dead internet” truly is. 

“As AI-generated content spreads, the challenge is finding a role for these models that doesn’t just result in a sanitized, repetitive web,” Dolezal said. “Rather than forcing models to be perfectly compliant and agreeable, allowing them to have a more distinct personality or ‘friction’ might help them act as a creative partner rather than a replacement for human voice.”

View the full article