Proliferation of AI-Generated News Websites and Content Farms Across Multiple Languages Degrading Information Integrity

Scores of AI-generated news websites and content farms are producing low-quality, clickbait content in a variety of languages. They are reportedly spreading false information and degrading the quality of information available online. These sites often lack human oversight, feature repetitive language, and sometimes fabricate information, posing a threat to the credibility of online news sources.

Risk classification

• Primary risk domain: 3 Misinformation
• Primary risk subdomain: 3.2 Pollution of information ecosystem and loss of consensus reality

The incident involves the coordinated, large-scale generation of low-quality, AI-written articles across 49 websites, polluting search engines and social feeds, and degrading the overall information ecosystem.

Additional risk subdomains

• 6.3 Economic and cultural devaluation of human effort: The automated generation of clickbait articles to capture programmatic ad revenue directly devalues and diverts funding away from legitimate, human-authored journalism.

Causal factors

• Entity: Human
• Intent: Intentional
• Timing: Post-deployment

The creation and deployment of these AI-generated content farms was an intentional decision by human operators seeking to generate advertising revenue.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Limited Risk: The report describes AI-generated content being used on websites designed to mimic human communication without disclosing that the content is AI-generated, which violates transparency obligations for chatbots and synthetic content.

AI system and alleged parties

• AI system: ChatGPT, unspecified (OpenAI)
• AI purpose: Writing Assistant; Automated Content Curation
• Behaviour type: Tool
• Alleged developer: Unnamed, unknown
• Alleged deployer: WaveFunction.info, TNewsNetwork.com, ScoopEarth.com, Maria Spanadoris, Many, HistoryFact.in, HarmonyHustle.com, GetIntoKnowledge.com, FilthyLucre.com, Famadillo.com, CountyLocalNews.com, CelebritiesDeaths.com, BestBudgetUSA.com, Adesh Ingale
• Alleged harmed parties: Journalists, General public

Harm severity

Severity ratings are based on this scale.

Highest direct severity in any category: Minor. Severity is scored from Negligible to Catastrophic in each harm category, for harm the reports describe as caused directly or indirectly by the AI system.

• Physical: direct Negligible, indirect Negligible
• Infrastructure: direct Negligible, indirect Negligible
• Property: direct Negligible, indirect Negligible
• Financial: direct Negligible, indirect Minor
• Environmental: direct Negligible, indirect Negligible
• Malicious content: direct Minor, indirect Negligible
• Differential treatment: direct Negligible, indirect Negligible
• Civil rights: direct Negligible, indirect Negligible
• Democracy: direct Negligible, indirect Negligible
• Privacy: direct Negligible, indirect Negligible
• Psychological: direct Negligible, indirect Negligible
• Epistemic: direct Minor, indirect Minor
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Financial

Reported: The report describes how these sites are designed to generate revenue from programmatic ads, which finance much of the world's media, but does not quantify the exact financial losses.

Indirectly caused: Legitimate news organizations and media outlets lose advertising revenue to these automated content farms.

Inferred additional harm: It is highly likely that advertisers unintentionally spent thousands of dollars on programmatic ads served on these 49 low-quality AI-generated sites, and legitimate media outlets lost equivalent revenue.

Malicious content

Reported: The report explicitly describes some false narratives created directly by the incident, such as a fabricated article claiming Joe Biden had died.

Directly caused: The creation of false narratives, such as a fabricated article claiming Joe Biden had died, which was published on CelebritiesDeaths.com.

Inferred additional harm: It is likely that other false or misleading articles were generated across the 49 identified websites, potentially reaching thousands of readers.

Epistemic

Reported: The report explicitly describes epistemic harm through the creation of unreliable, AI-generated news sites that fabricate information and publish false narratives.

Directly caused: The generation of fabricated news articles, such as a false report of Joe Biden's death, and the creation of 49 websites publishing low-quality, AI-generated content mimicking real news.

Indirectly caused: The dilution of the online information ecosystem with low-quality, automated clickbait, making it harder for users to find reliable, human-authored news.

Inferred additional harm: Thousands of online news consumers were likely exposed to fabricated or low-quality information, eroding their ability to distinguish real news from AI-generated content.

People affected

• Occurrences reported: 1
• People reportedly exposed: 124000

Potential causes

Management

• Profit Over Content Quality: Operators prioritized ad revenue over journalistic ethics and accuracy.
• Concealment of AI Usage: Site owners denied using AI and hid behind fake human personas.

Technology

• Generative LLM Hallucinations: AI models fabricated false details, such as G7 economic decisions.
• Error Message Output Pass-through: AI refusals and system prompts were published directly without filtering.
• Automated Site Generation Tools: Tools automatically created fake About Us and policy pages.

Data Inputs

• Unverified Source Summarization: AI summarized or rewrote articles from CNN and other sources unchecked.
• Prompting for Misinformation: Users prompted AI to create fake news, like false celebrity deaths.
• Lack of Input Data Verification: No validation of the truthfulness of the input sources used by the AI.

Human Factors

• Zero Human Editorial Oversight: Operators published AI outputs directly without any human review or editing.
• Intentional Deception by Owners: Site owners used AI to generate fake author profiles and buy followers.
• Reader Deception: Unassuming readers could not easily distinguish AI text from human writing.

Process and Methods

• Programmatic Ad Monetization: Algorithmic ad placement incentivized high-volume clickbait generation.
• Bulk Automated Publishing: Sites published hundreds of rewritten articles daily using automated workflows.
• Ineffective Detection Methods: Post-hoc detection tools like GPTZero were not integrated into ad networks.

Regulatory Environment

• Lack of AI Content Regulations: No legal framework forced disclosure of AI-generated news content.
• Lack of Ad Network Oversight: Programmatic ad platforms failed to audit and block AI content farms.

Information quality

• Classification confidence: High
• Reason for confidence: The report by NewsGuard is highly detailed, providing specific examples of websites, languages, and AI-generated errors. The role of generative AI (specifically ChatGPT) is explicitly documented through error messages like 'as an AI language model' found on the sites.
• Ambiguities identified: The exact scale of financial revenue generated by these content farms is not quantified, and the exact number of readers/consumers is mostly inferred from social media follower counts.
• Alternative interpretations: None. The evidence of AI generation is undeniable due to the presence of system error messages in the published text.

The incident involves the discovery of 49 AI-generated websites operating as content farms to capture ad revenue. While primarily motivated by financial gain rather than state-sponsored information warfare, the automated generation of false narratives highlights how low-cost generative AI can scale misinformation, posing a minor but persistent threat to information integrity.

• Overall national security impact: Minor
• Scope: Multiple nations
• Primary target: No clear primary
• Other affected: Global
• Alleged perpetrator: Unknown

Threat characteristics

• Imminence: Long-term. The proliferation of AI-generated content farms represents an ongoing, strategic challenge to information integrity rather than an immediate national security crisis.
• Autonomy: Human-supervised. AI models generated and published text autonomously through automated pipelines, though human operators established the infrastructure and monetization systems.
• Novelty: Evolved capability. Automated content farms are an established threat, but the use of advanced LLMs to generate highly realistic, multi-lingual text at scale represents an evolved capability.

Impact by dimension

• Physical security: Negligible. The incident involves purely digital text generation for content farms with no physical systems, critical infrastructure, or kinetic capabilities affected.
• Information security: Minor. Generation of false narratives like Joe Biden's death constitutes automated disinformation, but the primary driver was financial gain rather than a coordinated state-sponsored campaign.
• Sovereignty: Minor. The fabrication of a US President's death could theoretically impact public trust, but there is no evidence of actual disruption to government operations or constitutional processes.
• Economic security: Negligible. Financial impact is limited to redirection of programmatic advertising revenue, posing no threat to national economic stability or strategic technology sectors.
• Societal stability: Minor. Low-quality information pollution and fake news can degrade public trust over time, but this specific incident caused no widespread civil unrest or systemic human rights violations.