将人工智能应用于废物管理中的职业与环境健康：一项系统综述。
完整摘要 · 中文译文
目的：本系统综述探讨人工智能（AI）如何减轻废物管理中的职业和环境健康风险，在该行业中，工人和社区面临传染性、毒性和物理性危害。 方法：遵循PRISMA指南，我们系统检索了SCOPUS数据库并审查了81篇同行评审文章。分析重点在于所应用的人工智能技术类型、其在废物管理各阶段的部署情况以及其所针对的具体健康风险。 结果：分析识别出人工智能应用所针对的五类健康风险。人工智能被部署于废物管理的五个阶段和四种技术范式（监督学习、无监督学习、强化学习和深度学习）。人工智能通过三种主要途径减轻危害：自动化风险源隔离、高风险任务的机器人替代以及宏观层面的暴露监测。 结论：当前的公共卫生研究在很大程度上忽视了废物管理技术创新带来的健康风险降低收益。本文讨论了能够促进负责任的人工智能部署的制度创新。这种跨学科视角强调了跨部门合作解决公共卫生问题的必要性。
完整摘要 · 英文原文
OBJECTIVES: This systematic review examines how artificial intelligence (AI) mitigates occupational and environmental health risks in waste management, a sector where workers and communities face infectious, toxic, and physical hazards. METHODS: Following PRISMA guidelines, we systematically searched the SCOPUS database and reviewed 81 peer-reviewed articles. The analysis focused on the types of AI technologies applied, their deployment across waste management stages, and the specific health risks they address. RESULTS: The analysis identifies five types of health risks targeted by AI applications. AI is deployed across five waste management stages and four technical paradigms (supervised, unsupervised, reinforcement, and deep learning). AI alleviates hazards via three primary pathways: automated risk-source isolation, robotic substitution for high-risk tasks, and macro-level exposure monitoring. CONCLUSION: Current public health research has largely neglected health risk reduction gains from waste management technological innovation. Institutional innovations that can facilitate responsible AI deployment are discussed. This interdisciplinary perspective underscores the necessity of cross-sector collaboration to tackle public health issues.