Abdul Ghafur, Miki Nagao, Souha S Kanj, Salam Abbara, Pierre Tattevin, Mariette Awad, Maya Dagher, Linus K Ndegwa, Gathai Mundia, Jeroen Schouten, Huong T L Vu, Gabriel Levy Hara, Maneesh Paul Satyaseela, Mamoon A Aldeyab, Raksha K Bhat, Nitin Bansal, Heiman Fl Wertheim
AI should be viewed as a decision-support tool that augments, rather than replaces, AMS expertise. When implemented responsibly within robust governance frameworks and aligned with AMS principles, AI can contribute meaningfully to improved antimicrobial use and AMR control. Sustained human leadership, equity-focused deployment, and health system strengthening remain essential.
BACKGROUND: Antimicrobial resistance (AMR) continues to threaten the effectiveness of modern medicine and poses major public health challenges worldwide. Antimicrobial stewardship (AMS) programmes are essential to mitigate inappropriate antimicrobial use but are often resource-intensive and difficult to sustain. Advances in artificial intelligence (AI) have generated interest in its potential to support stewardship activities across clinical, operational, and public health domains.
OBJECTIVES: To critically review the current and emerging roles of AI in AMS, and to examine its opportunities, limitations, and implications for clinical practice, public health, and health systems globally.
METHODS: We conducted a narrative review of the literature on AI applications relevant to AMS, including diagnostic support, antimicrobial prescribing, behavioural and implementation science, infection prevention, surveillance, One Health approaches, drug discovery, and governance. Particular attention was paid to ethical considerations, human oversight, and applicability in low- and middle-income country (LMIC) settings.
RESULTS: AI has demonstrated potential to enhance diagnostic precision, support clinical decision-making, optimise antimicrobial use, strengthen surveillance systems, and inform targeted stewardship interventions. However, challenges related to data quality, bias, interpretability, workflow integration, ethical governance, and equity remain significant. AI systems developed in high-income settings may not be directly transferable to LMICs without careful adaptation and local validation.
CONCLUSIONS: AI should be viewed as a decision-support tool that augments, rather than replaces, AMS expertise. When implemented responsibly within robust governance frameworks and aligned with AMS principles, AI can contribute meaningfully to improved antimicrobial use and AMR control. Sustained human leadership, equity-focused deployment, and health system strengthening remain essential.