A Conversation with ChatGPT on Contentious Issues in Senescence and Cancer Research.

ABSTRACT

Artificial intelligence (AI) platforms, such as Generative Pretrained Transformer (ChatGPT), have achieved a high degree of popularity within the scientific community due to their utility in providing evidence-based reviews of the literature. However, the accuracy and reliability of the information output and the ability to provide critical analysis of the literature, especially with respect to highly controversial issues, has generally not been evaluated. In this work, we arranged a question/answer session with ChatGPT regarding several unresolved questions in the field of cancer research relating to therapy-induced senescence (TIS), including the topics of senescence reversibility, its connection to tumor dormancy, and the pharmacology of the newly emerging drug class of senolytics. ChatGPT generally provided responses consistent with the available literature, although occasionally overlooking essential components of the current understanding of the role of TIS in cancer biology and treatment. Although ChatGPT, and similar AI platforms, have utility in providing an accurate evidence-based review of the literature, their outputs should still be considered carefully, especially with respect to unresolved issues in tumor biology. SIGNIFICANCE STATEMENT Artificial Intelligence platforms have provided great utility for researchers to investigate biomedical literature in a prompt manner. However, several issues arise when it comes to certain unresolved biological questions, especially in the cancer field. This work provided a discussion with ChatGPT regarding some of the yet-to-be-fully-elucidated conundrums of the role of therapy-induced senescence in cancer treatment and highlights the strengths and weaknesses in utilizing such platforms for analyzing the scientific literature on this topic.