This letter warns about the identification of research publications available online that contain references to non-existent papers (ghost bibliographic references). This is likely to happen due to the use of ChatGPT, a LLM-based (Large-Language Models) chatbot which generates plausible but fake bibliographic references. Although the magnitude of the problem is still unknown (it could affect only preprint platforms, predatory journals, or poor-quality publication venues) and it is expected that ChatGPT would provide true references in a short period of time, both journals and publishers must be vigilant to prevent fake references from flooding their pages.