Stanford Study Reveals AI Now Powers Over 35% of New Websites—Is the 'Dead Internet' Theory Becoming Reality?

Conspiracy theorists have long speculated about the "dead internet" theory—the idea that online spaces, once dominated by human-created content, are increasingly filled with bots posing as people. While some allege this shift is deliberate, driven by governments or corporations, the rise of AI since ChatGPT’s 2022 debut has made the theory feel more plausible than ever.

Now, a groundbreaking study suggests the theory may be closer to reality than previously thought. Conducted by researchers at Stanford University, Imperial College London, and the Internet Archive, the study analyzed web pages published between 2022 and 2025 to determine how much online text is AI-generated. Using the Internet Archive’s Wayback Machine and multiple AI-detection methods, the team found that by May 2025, 35.3% of all new websites were either AI-generated or AI-assisted. Of those, 17.6% were created entirely by AI.

The study’s findings align with other recent data. Cloudflare reported that nearly a third of all internet traffic over the past year came from bots, while Imperva found that automated traffic surpassed human traffic for the first time in 2024.

"I find the sheer speed of the AI takeover of the web quite staggering. 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."

Jonáš Doležal, study researcher

AI-Generated Content: Defying Expectations

While AI-generated websites are becoming more common, the study suggests they may not be as harmful as critics fear. Researchers tested AI-generated content against six common critiques of AI text, confirming only two:

• Semantic contraction: A reduction in diverse viewpoints online.
• Positivity shift: Online writing is becoming more sanitized and artificially cheerful.

Surprisingly, the study found no evidence of other feared outcomes, including:

• Increases in rambling, low-substance text.
• A single generic writing style dominating the web.
• A lack of cited sources in AI-generated content.
• The spread of misinformation through AI-generated text.

The study’s authors, speaking to 404 Media, are developing a tool to continuously track AI-generated content, allowing users to monitor the "dead internet" phenomenon in real time.

"As AI-generated content spreads, the challenge is finding a role for these models that doesn’t just result in a sanitized,