Artificial intelligence in antimicrobial resistance surveillance: comparative analysis of global dashboards and predictive metrics

Background: Antimicrobial resistance poses an escalating global public health threat, with carbapenem-resistant Klebsiella pneumoniae rising 61% above European Union baselines and only 39% of National Action Plans fully financed. Major health agencies, including the WHO, ECDC, UKHSA, US CDC, Africa CDC, AAMRNet and Public Health Ontario, have developed surveillance systems and dashboards to monitor resistance trends and antimicrobial consumption. Artificial intelligence offers important potential for real-time, predictive AMR surveillance. Objectives: To evaluate antimicrobial resistance surveillance dashboards and metric frameworks across WHO, ECDC, UKHSA, US CDC, Africa CDC, PHO and AAMRNet, assess how AI could support decision-making, and identify policy gaps limiting predictive and actionable surveillance. Methods: A comparative policy analysis was conducted across seven global AMR surveillance systems using publicly available reports from 2022–2025. Frameworks were examined across dashboard purpose, reported metrics, data standardisation, interoperability and policy translation. AI readiness was assessed using criteria including data structure, timeliness, integration of microbiology and antimicrobial use datasets, and predictive analytics capacity. Data extraction followed a structured framework, with quantitative benchmarking and qualitative thematic analysis. Ethical approval was not required, as only publicly available aggregate data were used. Results: Surveillance maturity varied considerably. ECDC, UKHSA and US CDC emerged as the most advanced systems, with robust dashboard infrastructure, standardised resistance metrics and strong interoperability. MRSA bloodstream infections declined 20.4% below 2019 EU baselines, while laboratory participation increased, with 1,993 laboratories reporting to ECDC in 2024. Qualitative analysis identified four themes: strengthening surveillance infrastructure, expanding real-time monitoring, advancing antimicrobial stewardship programmes, and progressing One Health integration. Existing frameworks provide promising foundational datasets for AI-enabled predictive modelling, real-time genomic surveillance and decision-support tools. Conclusions: Global antimicrobial resistance surveillance systems are strengthening, with leading agencies demonstrating robust infrastructure, expanding dashboards and improving stewardship outcomes. These frameworks provide a strong foundation for AI integration, particularly where standardised, interoperable and longitudinal data exist. However, financing gaps, retrospective reporting limitations and One Health integration require further attention. Prioritising real-time genomic surveillance, harmonised data pipelines and AI-enabled predictive modelling can transform surveillance into actionable public health intelligence.