The study tested six hypotheses about AI content's effects on web quality. It confirmed two: semantic contraction, meaning reduced diversity of viewpoints, and a positivity shift toward more sanitized, cheerful language. The researchers found no evidence supporting concerns about rambling text, generic style, missing citations, or increased misinformation. The full scope of the study's methodology and additional findings remain under review.
The findings validate elements of the "dead internet" theory, which emerged around 2016 and posits that bot and AI dominance erodes authentic human interaction. Recent data supports the underlying concern: Cloudflare reported that nearly a third of web traffic now originates from bots, while Imperva documented automated traffic surpassing human traffic in 2024. For attorneys tracking AI liability, content authenticity, and platform governance issues, the study's continuous monitoring tool—which researchers plan to deploy—will provide ongoing benchmarks for how AI-generated content reshapes the information landscape.