The authors analyze the materials in librarianship generated with artificial intelligence (AI). They demonstrate that AI generates the texts very close to the common style and format of publications in the area. The goal of the article is to warn the users of the AI-generated works and possibly, to prevent violation of publishing ethics. The authors undertook comprehensive comparison to evaluate the quality and similarity of text materials, and application of the generated predictive text (GPT) technology in librarianship. The advantages and disadvantages of both approaches are revealed. The findings of the study are important to understand the AI capabilities and limitations in librarianship and to identify the contexts where the computerized systems are most efficient as compared to human authors. The ChatGPT is an excellent reference and complementary tool based on the accumulated samples. The materials created by ChatGPT are legally close to encyclopedic or dictionary entries, however, they have no actual author(s), i.e. these materials are independent creative works therefore they can hardly be protected by general copyright. The authors recommend to indicate the AIgenerated parts of works, argue that further development of juridical foundations for GPT materials is needed, and invite everyone to discussion. The study is accomplished within the framework of the NRU HSE Fundamental Studies Program.