Ethics Core (FABRIC)

道德核心 (FABRIC)

• 批准号: 10473062
• 负责人: Bradley A. Malin
• 金额: $ 54.3万
• 依托单位: UNIVERSITY OF CALIFORNIA, SAN DIEGO
• 依托单位国家: 美国
• 财政年份: 2022
• 资助国家: 美国
• 起止时间: 2022-07-08 至 2022-12-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/10473062
• 关键词: Address; Algorithms; Area; Artificial Intelligence; Behavioral Research; Benefits and Risks; Bioethics; Biomedical Research; Bridge to Artificial Intelligence; Catalogs; Climacteric; Clinical; Collaborations; Collection; Communities; Consent; Data; Data Collection; Data Protection; Data Set; Development; Economics; Ensure; Environment; Ethical Issues; Ethics; European; Foundations; Future; Generations; Guidelines; Health Care Research; Healthcare; Human; Individual; Informatics; Infrastructure; Knowledge; Laws; Learning; Literature; Machine Learning; Medicine; Persons; Privacy; Process; Public Policy; Public Relations; Reflex action; Regulation; Research; Research Ethics; Rights; Running; Societies; Surveys; Sustainable Development; Technology; Time; Trust; Unconscious State; Vision; Work; World Health Organization; adverse outcome; clinical decision-making; community engagement; data management; decision making algorithm; digital; digital health; distrust; experience; health care delivery; human centered design; implementation science; improved; interdisciplinary approach; legal implication; multidisciplinary; outreach; prevent; programs; racial bias; response; scaffold; social; success; symposium; technology development; tool; trustworthiness; usability; willingness

Bridge2AI: a FAIR AI BRIDGE Center (FABRIC) Ethics Core Summary The use of artificial intelligence (AI), and particularly machine learning (ML), in healthcare opens up many opportunities to improve healthcare and biomedical research. However, AI/ML also raise important issues that implicate ethics and trust, including defining parameters for consent and re-use of personal data, protecting privacy, ensuring transparency and engagement with stakeholders about this research, and developing and deploying tools that are useful and valid for all people. Without an ethically robust set of principles and practices that are generalizable and reusable in a wide range of biomedical environments, AI/ML could violate personal rights, widen the gap between fairness and equality, and fan the flames of mistrust, as exemplified by recent work showing how racial bias can influence clinical decision algorithms. Our vision for the FAIR AI Bridge Center - Ethics Core (FABRIC-Ethics) is to ensure that AI/ML is developed and applied in an ethical and trustworthy manner. FABRIC-Ethics will support the Bridge2AI program to become sustainable by making it more ethical and trustworthy by the end of the four-year project period. To realize this vision, we will use an iterative and reflective four-step cycle: 1) Scaffold, 2) Assess, 3) Facilitate and 4) Evaluate and educate, or SAFE, to provide a platform for convening, analyzing and curating, public relations and original research in a multidisciplinary manner. We will work with the Bridge2AI program to formulate ethical and trustworthy principles for AI/ML (ETAI) to address existing and future practices in biomedical AI research and applications. These include the collection and management of data, the development and deployment of AI/ML technologies and AI/ML applications. In close collaboration with the Bridge2AI program and its Data Generation Projects (DGPs), we will conduct a closed- and open-ended survey, discuss priorities and experiences with Bridge2AI DGPs, and develop an open, curated catalog of relevant literature. These efforts will run in parallel with multiple mechanisms for building a learning ETAI community, convening Bridge2AI data generation projects to distill best practices, and organizing studio sessions to support contact with the other core areas of the Bridge2AI Center and the broader community. Our core will further develop a digital health checklist and framework that prepares Bridge2AI DGPs to evaluate: 1) access and usability, 2) risks and benefits, 3) privacy and 4) data management. We will work with the Bridge2AI DGPs to share knowledge about ETAI, inform the development of principles and best practices, and to set up conferences for sustainable development of ETAI culture beyond Bridge2AI. The team assembled for the core has expertise in a wide range of areas, including bioethics, digital health research ethics, law, public policy, AI/ML, data protection, informatics, medicine, human- centered design, implementation science, and community engagement. To ensure success, FABRIC-Ethics will be led by four PIs with a proven track record in multidisciplinary approaches to the study of ethical issues in technology, center management, and core support.

Bridge2AI：一个公平的AI桥接中心(Fabric)伦理核心摘要 人工智能(AI)，特别是机器学习(ML)在医疗保健中的使用开启了许多 改善医疗保健和生物医学研究的机会。然而，AI/ML也提出了一些重要问题 牵涉到道德和信任，包括定义同意和重复使用个人数据的参数，保护 隐私，确保这项研究的透明度和与利益相关者的接触，并制定和 部署对所有人都有用和有效的工具。如果没有一套道德上可靠的原则和做法 在广泛的生物医学环境中可推广和可重用的，AI/ML可能会侵犯个人 权利，扩大公平和平等之间的差距，并煽动不信任的火焰，最近的例子就是如此 展示种族偏见如何影响临床决策算法的工作。我们对博览会AI桥梁中心的愿景 -伦理核心(织物-伦理)是确保人工智能/ML在伦理和值得信赖的环境中开发和应用 举止。布料伦理学将支持Bridge2AI计划，使其更加符合伦理道德，从而变得可持续 在四年项目期结束时值得信赖。 为了实现这一愿景，我们将使用迭代和反思的四步循环：1)脚手架，2)评估，3)促进 以及4)评估和教育，或安全，为召集、分析和策划公众提供一个平台 多学科的关系和原创研究。我们将与Bridge2AI计划合作 为AI/ML(ETAI)制定道德和值得信赖的原则，以解决现有和未来的实践 生物医学人工智能研究与应用。其中包括数据的收集和管理、开发 以及AI/ML技术和AI/ML应用的部署。与Bridge2AI计划密切合作 和它的数据生成项目(DGPS)，我们将进行封闭式和开放式调查，讨论优先事项 和使用Bridge2AI DGPS的经验，并开发一个开放的、经过策划的相关文献目录。这些努力 将与构建学习型ETAI社区的多个机制并行运行，召集Bridge2AI数据 生成项目以提炼最佳实践，并组织工作室会议以支持与其他核心的联系 Bridge2AI中心和更广泛的社区的区域。我们的核心将进一步开发一份数字健康清单 以及为Bridge2AI DGPS做好评估准备的框架：1)接入和可用性，2)风险和收益，3) 隐私和4)数据管理。我们将与Bridge2AI DGPS合作，分享关于EtaI的知识，通知 制定原则和最佳做法，并举办可持续发展会议 超越Bridge2AI的文化为核心组建的团队在广泛的领域拥有专业知识，包括 生物伦理学、数字健康研究伦理、法律、公共政策、人工智能/ML、数据保护、信息学、医学、人类- 以设计、实施科学和社区参与为中心。为了确保成功，织物-伦理将 由四名私人助理领导，他们在多学科方法研究道德问题方面有良好的记录 技术、中心管理和核心支持。