Why Do Chatbots Keep Telling Stories About Someone Named ‘Elias Thorne’? - Gizmodo
Your chatbot might be hallucinating the same fictional character. Here's why that matters for AI reliability.
Why it matters
This investigates a systemic pattern in LLM outputs—repeated fabrication of the same fictional entity across multiple models—which raises questions about training data contamination, memorization, and the reliability of AI systems in production environments.
The key facts
9 to knowMultiple chatbots generating identical or similar fictional narratives about 'Elias Thorne'
Suggests potential shared training data contamination or coordinated hallucination patterns
Raises reliability concerns for enterprises deploying chatbots in customer-facing roles
Published June 2026 — indicates ongoing AI quality/safety concerns persist
Multiple chatbots producing identical or near-identical stories about 'Elias Thorne' character
Pattern suggests training data contamination or shared data source problem
Indicates potential hallucination or false memory clustering across different AI systems
Raises concerns about model evaluation and the detectability of corrupted training data
Published June 2026—timeframe suggests this is a documented research or investigative finding
Go to the source
Reuters Technologynews.google.com
Publisher excerpt: Why Do Chatbots Keep Telling Stories About Someone Named ‘Elias Thorne’? Gizmodo