Study Finds A Third of New Websites Are AI-Generated

Why It Matters

The rapid infusion of AI‑written pages reshapes online information ecosystems, affecting brand credibility, SEO strategies, and the overall quality of public discourse. Understanding these shifts helps businesses and regulators anticipate risks and opportunities in a web increasingly populated by machine‑generated content.

Key Takeaways

• AI‑generated sites reached 35% of new domains by mid‑2025
• AI content made web text more positive and less semantically diverse
• Fact‑checkers found no rise in verifiable false statements on AI sites
• Researchers plan continuous monitoring via the Internet Archive’s Wayback Machine

Pulse Analysis

The surge of AI‑generated websites marks a watershed moment for the internet’s content supply chain. By mining Wayback Machine snapshots and applying the high‑precision Pangram v3 detector, Stanford and Imperial College scholars quantified the pace at which large‑language models have moved from novelty to mainstream publishing tools. Their methodology—pairing automated detection with human fact‑checking—offers a replicable blueprint for future audits of digital text, highlighting how quickly AI can dominate new domain registrations once a powerful model like ChatGPT becomes publicly accessible.

Beyond sheer volume, the study uncovers subtle shifts in tone and lexical richness. AI‑written pages tend to adopt a more cheerful, streamlined voice, reducing semantic density and narrowing stylistic variety. While this homogenization could simplify content consumption for some audiences, it also risks eroding the nuanced perspectives that fuel innovation and critical debate. Notably, the researchers did not observe a spike in outright falsehoods, suggesting that current AI systems, when guided by existing data, may prioritize plausibility over deception. However, the possibility of unverifiable claims slipping through underscores the need for robust verification pipelines, especially for sectors like finance, health and legal services where accuracy is paramount.

Looking ahead, the team’s partnership with the Internet Archive aims to create a live dashboard that flags AI‑generated trends in real time. Such a tool could become indispensable for marketers, SEO professionals, and policy makers seeking to differentiate authentic human‑crafted content from algorithmic output. As AI continues to lower the barrier to entry for web publishing, businesses must adapt their content strategies, invest in provenance tracking, and consider how to embed distinctive brand voice into a landscape that increasingly favors uniform, AI‑friendly prose.