This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
Bankrate and CNET, owned by Red Ventures, resumed publishing AI-generated articles that they claimed were human fact-checked. Investigations revealed that these articles contained significant factual errors, including incorrect economic and housing statistics. The company attributed these errors to the use of outdated datasets rather than the generative AI itself, despite marketing the content as authoritative and accurate.
Bankrate, and its sister site CNET, both owned by Red Ventures, resumed publishing AI-generated articles claiming thorough human fact-checking. However, new articles are alleged to contain numerous factual errors, including inaccurate statistics and misleading information. Despite public criticism, the company defended its use of AI and blamed out-of-date datasets for the errors. In addition to the errors, the incident raises questions about the ethical use of AI in journalism, especially given the company's insistence on "fact-checked" content.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
AI systems that inadvertently generate or spread incorrect or deceptive information, which can lead to inaccurate beliefs in users and undermine their autonomy. Humans that make decisions based on false beliefs can experience physical, emotional or material harms
Also classified as
Human editors over-relied on the AI system and failed to perform the promised thorough fact-checking before publication.
The AI system failed to perform reliably by generating inaccurate data points, demonstrating a lack of capability in factual generation.
The AI system generated and published false economic and housing statistics, spreading incorrect information to readers seeking financial advice.
AI system
Due to a decision or action made by an AI system
Unintentional
Due to an unexpected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed
The factual errors were generated by the AI system during its active deployment on Bankrate's live website, representing an unintended failure of the generation process.
Added categories peak at 2: Minor (Epistemic Harm, Direct)
The incident involved Bankrate publishing AI-generated financial articles containing minor factual and statistical errors due to outdated datasets and inadequate human oversight. This commercial publishing error poses negligible threats to national security, critical infrastructure, or sovereignty.
The AI system is described as an 'AI-assist tool' used to generate draft articles that are intended to be thoroughly edited and fact-checked by human editors before publication.
Risk Level 3: The system generates public-facing content, which falls under transparency obligations requiring users to be informed that the content is AI-generated.
The system's primary function is to generate written articles to assist the editorial team in publishing content quickly.
The report provides highly specific, verifiable examples of the AI's factual errors, comparing them directly to the cited source databases. The publisher's response and explanation are also documented, leaving little ambiguity about the nature of the failure.
The publisher argues this was a database retrieval error rather than a generative AI failure, though the output was still presented as an AI-generated article.
Technology
Flawed Database Integration
AI tool pulled outdated data from internal database without validation.
Lack of Automated Validation
No system to cross-reference generated text with live external sources.
Data Inputs
Outdated Internal Dataset
The database used for the article contained outdated data from last year.
Inaccurate Sourced Data
Data points for home prices and salaries did not match actual sources.
Human Factors
Inadequate Editor Review
Editors failed to catch blatant factual errors in the AI-generated text.
Overreliance on Automation
Staff assumed the AI-assist tool and database outputs were accurate.
Process and Methods
Flawed Fact-Checking Process
Verification workflow did not require validation against primary sources.
Hasty Publishing Workflow
Process prioritized rapid output over thorough content verification.
Management
Prioritizing SEO over Accuracy
Management pushed AI content to capture search traffic without safeguards.
Premature AI Resumption
Management restarted AI publishing before establishing robust reviews.
AI system involved
Identified by automated extraction; can differ from the alleged developers above.
