Abstract
Large Language Models (LLMs) have significantly shaped working practices across a variety of fields including academia. Demonstrating a remarkable versatility, these models can generate responses to prompts with information in the form of text, documents, and images, show ability to summarize documents, perform literature searches, and even more, understand human behavior. However, despite providing many clear benefits, barriers remain toward their integration into academic work. Ethical and practical concerns regarding their suitability for various tasks further complicate their appropriate use. Here, we summarize recent advances assessing the capacity of LLMs for different components of academic research and teaching, focusing on three key areas in the psychological sciences: education and assessment, academic writing, and simulating human behavior. We discuss how LLMs can be used to aid each area, describe current challenges and good practices, and propose future directions. In doing so, we aim to increase the awareness and proper use of LLMs in various components of academic work, which will only feature more heavily over time.