This chapter explains how PhD students can use large language models (LLMs) to support research without outsourcing scholarship. It covers practical uses in writing, literature exploration, coding, experiment planning, and self-critique, while emphasizing that you remain responsible for accuracy, reasoning, and originality. The chapter highlights common failure modes such as fabricated citations, confident errors, and overly agreeable “reviews,” and discusses when LLM feedback is helpful versus misleading. It also outlines responsible use, including disclosure norms, bias awareness, and protecting confidential data. The goal is to benefit from speed and breadth while staying intellectually in control.