A new study published in the peer-reviewed journal JMIR Mental Health by JMIR Publications highlights a critical risk in the growing use of Large Language Models (LLMs) like GPT-4o by researchers: the frequent fabrication and inaccuracy of bibliographic citations.
