Two independent Nature Biotechnology reports demonstrated deep‑learning approaches that accelerate antibacterial discovery by prioritizing novel chemical scaffolds from high‑throughput and phenotypic screens. One study combined phenotypic screening with deep learning to identify active compounds; the other trained models on large experimental datasets to surface structurally novel antimicrobials. Authors showed that AI models can expand chemical space beyond human intuition and recover hits missed by conventional methods. Both papers underscore the growing role of machine learning in addressing antimicrobial resistance by speeding hit discovery and de‑risking lead identification for notoriously difficult targets.