Human researchers outperformed large language models across all major stages of systematic review preparation, particularly in study selection, synthesis, and final manuscript drafting. While LLMs ...