Verify any claim · lenz.io
Claim analyzed
Tech“More than 50% of online content is generated by artificial intelligence rather than written by humans.”
Submitted by Vicky
The conclusion
The available evidence does not support the statement that most online content is AI-generated. The strongest broad estimate cited is below 50%, while the higher numbers refer to narrower categories such as newly published pages, English-language articles, pages containing some AI text, or automated traffic rather than human-versus-AI authorship. That makes the claim an overstatement of what current evidence shows.
Caveats
- Low confidence conclusion.
- Figures about bot traffic are not evidence that most content was written by AI; traffic volume and content authorship are different measures.
- Numbers for new pages or articles cannot be generalized to all online content, which includes older material, other formats, and non-English content.
- Some cited high percentages rely on detector-based or secondary reports and may blur 'contains AI text,' 'AI-assisted,' and 'fully AI-generated.'
Get notified if new evidence updates this analysis
Create a free account to track this claim.
Sources
Sources used in the analysis
Europol and other analysts have warned that up to 90 percent of online content may be synthetically generated by 2026. This is a future projection, not a current measurement.
The findings suggest that at least 30% of text on active web pages originates from AI-generated sources, with the actual proportion likely approaching 40%. While it is increasingly evident that the internet is becoming saturated with content created by generated AI large language models, accurately measuring the scale of this phenomenon has proven challenging.
Automated bot traffic surpassed human-generated traffic for the first time in a decade, constituting 51% of all web traffic in 2024. Malicious bots now account for 37% of all internet traffic, a significant increase from 32% in 2023.
In 2024, most of the global website traffic was still generated by humans, but bot traffic is constantly growing. Fraudulent traffic through bad bot actors accounted for 32 percent of global web traffic in the most recently measured period.
The latest data from the OECD’s AI Incidents and Hazard Monitor reveals a boom in AI-related content incidents to nearly 500 by January 2026. This underscores the rapid proliferation of AI-generated content worldwide, from synthetic media to deepfakes, but does not quantify the percentage share of all online content.
As the human-AI interactions generate less unique content, the more it may contribute to content homogenization and bias, the study suggests. A working paper by UCLA Anderson’s Francisco Castro and Jian Gao, a Ph.D. student, and Northwestern’s Sébastien Martin takes a human-centered approach to the problems of content homogenization and bias.
This paper studies the impact of generative AI on U.S. households' task allocation at home, using detailed Internet browsing data from a large sample of households. It does not quantify the proportion of online content generated by AI.
We analyzed 900,000 newly created web pages in April 2025 and found that 74.2% of them contained AI-generated content. 2.5% of pages were categorized as 'pure AI,' 25.8% as 'pure human,' and 71.7% as a mix. When surveyed, 87% of content marketers reported using AI to create or help create content.
In November 2024, the quantity of AI-generated articles being published on the web surpassed the quantity of human-written articles. The study classifies an article as AI-generated if the algorithm predicts that more than 50% of the content is AI-generated, using Surfer's AI detector with a chunk size of 500 words. The proportion of AI-generated articles has remained relatively stable over the last 12 months.
Using Common Crawl data, Graphite found that AI-generated writing had passed the 50% mark of newly published web articles in November of last year. That figure has plateaued in recent months, but it's still a huge change in how content is produced. More new articles online are written by artificial intelligence than by human beings, according to a new study from Graphite.
New research from Graphite found that around half of all articles on the internet are AI generated. The analysis of 65,000 English-language articles from 2020-2025 showed AI-generated articles at 52% of new articles as of May 2025, after peaking in November 2024. However, 86% of articles in Google Search were human-written.
According to recent reports, automated traffic has surpassed human activity, accounting for 51% of all web traffic. Bad bots, specifically, make up a significant portion of this automated traffic.
AI content detection tools have known limitations, including false positive rates (Graphite.io reported 4.2% for SurferSEO's detector) and difficulty distinguishing between AI-generated, AI-assisted, and human-written content. Most studies measure published articles rather than total online content, which includes images, videos, and other media formats not captured in article-focused analyses.
Ahrefs analyzed nearly a million new web pages published in April 2025 and found that 74.2 percent contained detectable AI generated content. Graphite, an SEO firm, examined more than 60,000 new articles from 2020 through 2025 and found that by late 2024, more than half of new English language articles were primarily AI written. That percentage grew into 2025. One widely cited analysis estimates that about 57 percent of all online text has been generated or translated using AI tools.
One widely cited analysis estimates that about 57 percent of all online text has been generated or translated using AI tools. This includes machine translation, rewriting, and generative writing.
What do you think of the claim?
Your challenge will appear immediately.
Challenge submitted!
Expert review
How each expert evaluated the evidence and arguments
Expert 1 — The Logic Examiner
The claim asserts that "more than 50% of online content is generated by AI rather than written by humans," but the evidence pool reveals critical inferential gaps: the most direct broad measurement (Source 2, arXiv) places AI-origin text at 30–40% of active web pages, explicitly below the 50% threshold; Sources 9–11 (Graphite/TechRadar/Futurism) only demonstrate that newly published English-language articles crossed 50%, which is a narrow temporal and categorical subset of "all online content"; Source 8 (Ahrefs) shows 74.2% of new pages contain some AI content but conflates AI-assisted with AI-generated; Source 3 (Thales/Imperva) measures bot traffic requests, not authored content; and the "57% of all online text" figure (Sources 14–15) appears in low-authority commentary without traceable primary measurement. The Opponent's rebuttal correctly identifies the composition fallacy (equivocating from "new articles" to "all online content"), the false equivalence between bot traffic and AI-authored content, and the reliance on unverified secondary citations, while the Proponent's rebuttal fails to close the inferential gap between newly published article majorities and the total stock of online content — the claim as stated is therefore not logically supported by the evidence and is best classified as Misleading.
Expert 2 — The Context Analyst
The claim blurs key categories by treating findings about “new articles/new pages” and “pages containing some AI text” (Sources 8–11) and even bot traffic shares (Source 3) as if they measured the share of all existing online content that is AI-written; it also omits that the most direct broad estimate in the pool puts AI-origin text on active web pages at ~30–40% and highlights measurement uncertainty (Source 2), while Europol's high figures are explicitly projections (Source 1). With the full context restored, the evidence does not support that a current majority of overall online content is AI-generated rather than human-written, so the claim gives a misleading overall impression and is effectively false.
Expert 3 — The Source Auditor
The highest-authority, most directly relevant sources in the pool do not establish a current >50% share of all online content being AI-generated: the Europol item (Source 1, Europol) is explicitly a future projection, and the only broad quantification study cited (Source 2, arXiv preprint) estimates ~30–40% of text on active web pages is AI-origin and stresses measurement difficulty; the Imperva/Thales and Statista items (Sources 3–4) address bot/traffic rather than authored content. The main >50% support comes from less independent, detector-based or subset-specific reporting (Sources 8–11, Ahrefs/Graphite/TechRadar/Futurism) about new pages or articles (often English-only) and is amplified by low-authority secondary repetition (Sources 14–15), so trustworthy evidence in this brief fails to support the claim and points below 50% for overall web text.
Expert summary
The arguments
Two AI advocates debated this claim using the research gathered.
Argument for
Multiple independent studies converge on the conclusion that AI-generated content has crossed the 50% threshold: Source 9 (Graphite.io) and Source 10 (TechRadar) both confirm that AI-generated articles surpassed human-written articles in November 2024 using Common Crawl data, Source 11 (Futurism) reports AI articles at 52% of new content as of May 2025, Source 8 (Ahrefs) found 74.2% of newly created web pages contained AI-generated content, and Source 14 (MyNewITGuys) cites a widely referenced analysis estimating 57% of all online text has been generated or translated using AI tools. Furthermore, Source 3 (Thales Group / Imperva) establishes that automated traffic already constitutes 51% of all web traffic, and Source 2 (arXiv) — despite its refuting stance — concedes that at least 30–40% of active web page text is AI-sourced, a figure that, combined with the article-level majority documented by Graphite and Ahrefs, strongly supports the claim that AI-generated content now exceeds the 50% mark across online content broadly.
The Proponent's case equivocates from “new articles/pages” to “all online content”: Graphite/TechRadar/Futurism address a narrow slice (English-language articles and newly published items) and rely on detector-based classifications that blur AI-generated vs AI-assisted text (Sources 9–11, 13), while Ahrefs only finds that 74.2% of new pages contain some AI content—not that a majority of total content is AI-written (Source 8). The Proponent then compounds this category error by treating bot traffic as authored content (Source 3) and by leaning on an unsubstantiated “57% of all online text” claim repeated in low-authority commentary/discussion rather than measured evidence, whereas the only direct broad quantification in the brief estimates ~30–40% AI-origin text on active pages and explicitly flags measurement limits—still below the motion's >50% threshold (Sources 2, 14–15).
Argument against
The motion claims a current majority of all online content is AI-generated, but the most direct quantification in the brief puts AI-origin text on active web pages at ~30–40%, explicitly noting measurement difficulty and falling well short of 50% (Source 2, arXiv: “Delving into… quantification of AI-generated content”). The main “over-50%” evidence is either a future projection (Source 1, Europol) or narrowly about subsets like new articles/pages and even then relies on imperfect detectors and mixed/AI-assisted definitions (Sources 8–11, Ahrefs/Graphite/TechRadar/Futurism; Source 13), while bot-traffic figures (Source 3, Thales/Imperva) measure requests not authored content—so none substantiates that >50% of overall online content is AI-written.
The Opponent's reliance on Source 2 (arXiv) as the definitive ceiling is a selective reading fallacy: Source 2 explicitly acknowledges measurement difficulty and concedes the figure is "likely approaching 40%" — a lower-bound estimate on active pages that does not account for the surge in newly published content documented by Source 8 (Ahrefs) and Source 9 (Graphite.io), where AI-generated articles demonstrably crossed 50% by November 2024. Furthermore, the Opponent's dismissal of Sources 8–11 as "narrow subsets" ignores that newly published articles and web pages constitute the very frontier of online content growth, and Source 11 (Futurism) — drawing on a 65,000-article longitudinal dataset spanning 2020–2025 — is precisely the kind of broad, time-series evidence that substantiates a majority claim, not merely a marginal or projected one.