Brain foundation models bring the foundation model paradigm to the field of neuroscience. Like language and image foundation models, they are general-purpose AI systems pretrained on large-scale datasets that adapt readily to downstream tasks. Unlike text-and-image based models, however, they train on brain data: large-datasets of EEG, fMRI, and other neural data types historically collected within tightly governed clinical and research settings. This paper contends that training foundation models on neural data opens new normative territory. Neural data carry stronger expectations of, and claims to, protection than text or images, given their body-derived nature and historical governance within clinical and research settings. Yet the foundation model paradigm subjects them to practices of large-scale repurposing, cross-context stitching, and open-ended downstream application. Furthermore, these practices are now accessible to a much broader range of actors, including commercial developers, against a backdrop of fragmented and unclear governance. To map this territory, we first describe brain foundation models' technical foundations and training-data ecosystem. We then draw on AI ethics, neuroethics, and bioethics to organize concerns across privacy, consent, bias, benefit sharing, and governance. For each, we propose both agenda-setting questions and baseline safeguards as the field matures.
Towards Adaptive Categories: Dimensional Governance for Agentic AI
The Digital Statecraft Academy（数字国家学府）；Department of Computer Science, University College London (UCL)（伦敦大学学院计算机系）；Department of Mathematics, Imperial College London（伦敦帝国学院数学系）
Digital health transformation in Quebec: assessment of interoperability and governance strategies
École de santé publique de l’Université de Montréal（蒙特利尔大学公共卫生学院）；Centre intégré universitaire de santé et de services sociaux du Centre-Sud-de-l'Île-de-Montréal（蒙特利尔岛中南地区综合大学卫生与社会服务服务中心）；Centre interuniversitaire de recherche en analyse des organisations（组织分析研究跨大学中心）
Open Problems in AI Risk Modeling: Insights from a Workshop on the Technical Foundations of AI Risk Modeling
Institute for Security and Technology（安全与技术研究所）；Center for Long-Term Cybersecurity（长期网络安全中心）；SaferAI；Center for Security and Emerging Technology（新兴技术安全中心）；University of Warwick（华威大学）；Center for AI Risk Management & Alignment（人工智能风险管理与对齐中心）
Artificial Intelligence and Structural Injustice: Foundations for Equity, Values, and Responsibility
本章主张采用结构性不公正的方法来治理人工智能。结构性不公正具有分析性和评价性成分。分析部分包括社会科学中众所周知的结构性解释。评价的成分是一种正义理论。结构性不公正是一个强大的概念工具，允许研究人员和从业者识别，表达，甚至预测人工智能偏见。本章以人工智能中的种族偏见为例，这种偏见源于结构性不公正。本章随后介绍了哲学家Iris Marion Young提出的结构性不公正的概念。本章还认为，结构性不公正非常适合作为人工智能治理的一种方法，并将这种方法与从危害和利益分析或价值陈述开始的替代方法进行了比较。本章认为，结构性不公正提供了多样性，公平和包容的价值观和关注的方法和规范基础。本章最后展望了结构和责任的概念。结构的概念是正义的核心。一个开放的理论研究问题是，人工智能本身在多大程度上是社会结构的一部分。最后，责任的实践是结构性不公正的核心。即使他们不能对结构性不公正的存在负责，每个人和每个组织都有责任解决未来的结构性不公正。
Governing Well in the Algorithmic Age: The Foundations of Digital Statecraft
The Digital Statecraft Academy（数字治国研究院）；UCL（伦敦大学学院）；University of Cambridge（剑桥大学）；University of Bristol（布里斯托大学）；Department for Digital, Culture, Media and Sport (DCMS), UK Government（英国政府数字、文化、媒体和体育部）；The Alan Turing Institute（艾伦·图灵研究所）；Swiss Federal Office of Energy（瑞士联邦能源办公室）；Imperial College London（帝国理工学院）；UK Statistics Authority（英国统计局）；Department of Labor, State of Maryland（马里兰州劳工部）；Yale University（耶鲁大学）；Innovorsa Group（Innovorsa集团）；Dataswyft（Dataswyft公司）；University of Tartu（塔尔图大学）；Prime Minister’s Office, UAE（阿联酋总理办公室）；Hitotsubashi University（一桥大学）；University of Oxford（牛津大学）；James Madison University（詹姆斯·麦迪逊大学）；NYU GovLab（纽约大学治理实验室）；The Data Tank（数据智库）
