A new research paper from Samsung and the University of Warsaw has identified a phenomenon where large language models repeatedly generate the same fictional expert names across various contexts. Names like Elena Vasquez and Marcus Chen have appeared as co-authors on hundreds of AI-generated documents, including academic papers and books, despite the individuals not existing. This research highlights how LLMs can AI
IMPACT Highlights potential for AI-generated content to be contaminated by recurring fictional personas, impacting academic publishing and creative works.
RANK_REASON Research paper detailing a phenomenon in LLM-generated content.
Read on Mastodon — mastodon.social →
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →