This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
G/O Media began publishing AI-generated articles using Google's Bard and OpenAI's ChatGPT despite strong objections from its editorial staff. The articles contained factual errors, such as a non-chronological list of Star Wars movies, which staff members argued damaged their professional credibility and journalistic standards. The incident highlights tensions between management's push for AI adoption and staff concerns regarding quality, ethics, and workload.
G/O Media began publishing AI-generated articles, against staff advice, that contained errors and quality issues. The first such article, a list of Star Wars movies, failed to maintain chronological order, causing internal concerns over journalistic credibility and ethics. Staff expressed that the AI was "actively hurting our reputations and credibility" and accused management of "wasting everyone's time."
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 deployment of AI writers devalues the professional efforts and reputations of the human journalists employed at the company.
The AI system generated a non-chronological list of Star Wars movies, which constitutes false or misleading information published as factual content.
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 immediate risk of reputational damage was triggered by the AI system generating inaccurate content, specifically a non-chronological Star Wars list.
G/O Media's deployment of AI-generated articles containing factual errors caused internal labor disputes and minor reputational damage. This incident is a commercial quality-control and labor issue with negligible implications for national security.
The AI systems (Google Bard and ChatGPT) were used to generate draft articles, acting as writing assistants to produce content, though they were published without the required human editorial oversight.
Limited Risk: The report describes the use of chatbots (Google Bard and ChatGPT) to generate articles. Under the EU AI Act, AI systems that generate or manipulate content (like chatbots) fall under Limited Risk, requiring transparency obligations to inform users they are interacting with AI.
The AI systems were used to write articles for a media website, which directly aligns with the function of a Writing Assistant.
The report clearly outlines the deployment of Google Bard and ChatGPT by G/O Media, the specific factual errors produced (the Star Wars list), and the resulting backlash from staff and the union. There is little ambiguity about the events, the systems used, or the nature of the dispute.
None
Technology
Inaccurate Chatbot Outputs
Google's Bard and OpenAI's ChatGPT produced inaccurate, out-of-order lists.
Lack of Fact-Checking Capabilities
The AI bots lacked internal verification mechanisms to ensure factual accuracy.
Data Inputs
Unverified Training Data
AI relied on patterns in training data without real-time factual validation.
Human Factors
Staff Resistance and Dismay
Journalists strongly objected to AI use, citing damage to their reputation.
Lack of Editorial Review
Editors did not review or approve the AI-generated content before publication.
Process and Methods
Bypassing Editorial Workflows
AI articles were published directly without standard editorial review processes.
Undefined Editing Roles
Job descriptions did not include reviewing or editing AI-produced content.
Management
Prioritizing AI Over Quality
Leadership pushed AI adoption despite staff objections and obvious errors.
Inadequate Risk Assessment
Management failed to evaluate the reputational risks of publishing AI errors.
Regulatory Environment
Lack of AI Publishing Standards
No industry regulations governed the labeling or verification of AI content.
AI system involved
Identified by automated extraction; can differ from the alleged developers above.
