This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
A reporter at a Wyoming newspaper resigned after it was discovered he used generative AI to fabricate quotes and content in multiple news articles. The deception was uncovered by a competing journalist who noticed robotic phrasing and false information. The newspaper's leadership apologized, implemented new detection systems, and retracted the affected stories.
A reporter at the Cody Enterprise used AI to generate fake quotes and stories, including fabricating statements from Wyoming’s governor and other officials. The misuse of AI was uncovered when another journalist noticed robotic phrases and false information in the articles. The reporter resigned, and the newspaper is now implementing policies to prevent future incidents.
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
The reporter overrelied on and trusted the AI system to generate accurate journalistic content without verifying the quotes.
The AI system generated false quotes and fabricated information that was published in news articles, misleading readers and the community.
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 incident was caused by the generative AI system hallucinating and fabricating quotes, which was an unexpected outcome of the reporter using it to assist with writing.
Added categories peak at 2: Minor (Epistemic Harm, Direct)
A local reporter in Wyoming used generative AI to fabricate quotes from public officials, including the state governor, leading to his resignation and story retractions. The incident represents a minor information integrity issue with negligible national security implications, as it was localized, lacked adversarial intent, and was managed through standard editorial procedures.
The reporter used the generative AI chatbot as a writing assistant to help draft articles and generate quotes, which requires human oversight and validation before publishing.
Limited Risk: The incident involves a generative AI chatbot used to generate text and articles, which falls under the category of chatbots and AI-generated content requiring transparency obligations.
The AI system was primarily used as a writing assistant to help the reporter draft news articles and generate quotes.
The reports from multiple reputable news sources provide consistent, detailed accounts of the incident, including the specific number of fabricated articles, the reactions of the editors, and the admission of the reporter. The only minor detail missing is the specific name of the commercial chatbot used, which does not impact the classification.
None. The facts of the case are clear and admitted by the involved parties.
Technology
Chatbot Hallucination
Chatbots are programmed to answer even if the output is fabricated garbage.
Plausible but Spurious Output
AI generated highly believable but completely fabricated quotes.
Data Inputs
Unverified Prompt Inputs
Putting raw documents into AI without validating the resulting outputs.
Human Factors
Misplaced Trust in AI
A hurried rookie reporter trusted AI to generate accurate quotes.
Inexperience of Rookie Reporter
A novice reporter new to journalism did not understand the risks of AI.
Process and Methods
Absence of AI Policy
The newspaper lacked any policy or guidelines regarding AI-generated content.
Inadequate Editorial Review
The editor failed to catch robotic phrasing and fabricated quotes.
Lack of AI Detection Tools
No system was in place to scan and identify AI-written articles.
Management
Omission of AI Risk Assessment
Management assumed it was obvious that journalists would not use AI.
Deficient Pre-Employment Screening
AI usage and ethics were not discussed during the hiring process.
AI system involved
Identified by automated extraction; can differ from the alleged developers above.
