AIの「ゴースト著者」が学術出版を汚染
研究者らが、LLMが架空の専門家に同じ名前を生成し、学術記録を汚染していることを発見。学術アグリゲーターと検証プロセスに影響を与える。
researchers from samsung and the university of warsaw discovered that large language models, such as chatgpt, gemini, and claude, generate consistent, fictional names for experts in certain fields. these names, like elena vasquez and marcus chen, appear in hundreds of ai-generated academic papers, articles, and books. the researchers found 1,655 ghost-authored records on zenodo, a cern-operated repository, with fabricated publication dates. this contaminates academic records and affects scholarly aggregators like researchgate and google scholar. the researchers propose using these consistent names to determine the provenance of ai-generated content.