Unequal harvests: AI-assisted evidence map of trends and gaps in global farmer health research along SDG 3 priorities

INTRODUCTION: Ensuring the health of agricultural workers, the world's largest labour force, is key for sustainable food production and progress towards the Sustainable Development Goals (SDGs). METHODS: We conducted an artificial intelligence (AI)-assisted evidence map of research records on global farmer health published from 2015 to 2024. We searched bibliographic databases and screened titles/abstracts using SWIFT-Active Screener, a collaborative review platform that uses machine-learning prioritisation to rank records for human review. We retrieved 32 006 records. After manually screening 8533 records and stopping when the tool estimated ≥94% recall of relevant records, we included 1684 studies. We mapped research output by health topic category (non-communicable diseases (NCDs), communicable diseases, injuries and mental health) country income groups and alignment with SDG 3 targets. RESULTS: Despite 98% of the agricultural workforce living in low- and middle-income countries (LMICs), 52% of studies originate from high-income countries (HICs). Research focuses on NCDs (29%) and injuries (26%), with LMICs focusing on pesticide poisoning and HIC on accidents. Mental health emerges as a key topic in HICs, with the proportion of publications nearly doubling from 2021 to 2024 but remains underexplored in LMICs. Key gaps with high relevance to farming populations and climate change, such as heat-related illnesses, occupational injuries and musculoskeletal conditions, are not well represented in SDG 3 indicators. CONCLUSION: Our findings highlight urgent needs for a more equitable and comprehensive global research agenda that integrates agricultural worker health into sustainability frameworks beyond the SDG era, ensuring the resilience and well-being of food producers worldwide.