Core Viewpoint - The article discusses the urgent need for new antimicrobial peptides (AMPs) as promising alternatives to traditional antibiotics, particularly in the context of combating multidrug-resistant bacteria, highlighted by the WHO's ESKAPE list, with carbapenem-resistant Acinetobacter baumannii (CRAB) being a primary concern [2]. Group 1: Research Development - A research team from Shandong University published a study in Nature Microbiology on October 3, 2025, focusing on a generative artificial intelligence approach for discovering AMPs against multidrug-resistant bacteria [3]. - The study developed a pre-trained protein large language model (LLM) named ProteoGPT, which efficiently explores the AMP space to address clinical superbugs [4][10]. Group 2: Methodology and Findings - ProteoGPT was further developed into specialized sub-LLMs to create a workflow capable of rapidly screening millions of peptide sequences for strong antibacterial activity while minimizing cytotoxic risks [7]. - In vitro experiments showed that both discovered and generated AMPs exhibited low susceptibility to resistance development in CRAB and methicillin-resistant Staphylococcus aureus (MRSA) [8]. - In animal models, these AMPs demonstrated treatment effects comparable to or better than clinically used antibiotics without causing organ damage or disrupting gut microbiota [8].