Nutbeam Don, Milat Andrew J
School of Public Health, University of Sydney, Camperdown, NSW, Australia; and Sydney Health Partners, NSW, Australia.
Public Health Res Pract. 2025 Mar;35. doi: 10.1071/PU24001.
Objectives and importance of the study Applications of artificial intelligence (AI) platforms and technologies to healthcare have been widely promoted as offering revolutionary improvements and efficiencies in clinical practice and health services organisation. Practical applications of AI in public health are now emerging and receiving similar attention. This paper provides an overview of the issues and examples of research that help separate the potential from the hype. Methods Selective review and analysis of cross-section of relevant literature. Results Great potential exists for the use of AI in public health practice and research. This includes immediate applications in improving health education and communication directly with the public, as well as great potential for the productive use of generative AI through chatbots and virtual assistants in health communication. AI also has applications in disease surveillance and public health science, for example in improving epidemic and pandemic early warning systems, in synthetic data generation, in sequential decision-making in uncertain conditions (reinforcement learning) and in disease risk prediction. Most published research examining these and other applications is at a fairly early stage, making it difficult to separate the probable benefits from the hype. This research is undoubtedly demonstrating great potential but also identifying challenges, for example in the quality and relevance of health information being produced by generative AI; in access, trust and use of the technology by different populations; and in the practical application of AI to support disease surveillance and public health science. There are real risks that current access and patterns of use may exacerbate existing inequities in health and that the orientation towards the personalisation of health advice may divert attention away from underlying social and economic determinants of health. Conclusions Realising the potential of AI not only requires further research and experimentation but also careful consideration of its ethical implications and thoughtful regulation. This will ensure that advances in these technologies serve the best interests of individuals and communities worldwide and don't exacerbate existing health inequalities.
研究的目标与重要性 人工智能（AI）平台和技术在医疗保健领域的应用已得到广泛推广，被认为能在临床实践和卫生服务组织方面带来革命性的改善和效率提升。目前，AI在公共卫生领域的实际应用正在兴起并受到类似关注。本文概述了相关问题及研究实例，以帮助区分潜力与炒作。方法 对相关文献进行选择性综述与分析。结果 AI在公共卫生实践和研究中具有巨大潜力。这包括在直接改善健康教育以及与公众沟通方面的即时应用，以及通过聊天机器人和虚拟助手在健康传播中有效利用生成式AI的巨大潜力。AI还可应用于疾病监测和公共卫生科学，例如改进流行病和大流行早期预警系统、生成合成数据、在不确定条件下的序贯决策（强化学习）以及疾病风险预测。大多数已发表的关于这些及其他应用的研究尚处于相当早期阶段，难以区分可能的益处与炒作。这项研究无疑展示了巨大潜力，但也发现了挑战，例如生成式AI产生的健康信息的质量和相关性；不同人群对该技术的获取、信任和使用；以及AI在支持疾病监测和公共卫生科学方面的实际应用。当前的获取和使用模式存在实际风险，可能会加剧现有的健康不平等，而且对健康建议个性化的关注可能会转移对健康的潜在社会和经济决定因素的注意力。结论 要实现AI的潜力，不仅需要进一步的研究和试验，还需要仔细考虑其伦理影响并进行深思熟虑的监管。这将确保这些技术的进步符合全球个人和社区的最大利益，而不会加剧现有的健康不平等。
