Collaborative Research: SaTC: CORE: Medium: A Large-Scale, Longitudinal Resource to Advance Technical and Legal Understanding of Textual Privacy Information

消费者隐私保护的一个关键方面是透明度。然而，尽管有大量与组织隐私实践相关的隐私政策和其他文本，但分析这些文件一直是大规模理解数字隐私的障碍。利用隐私政策、服务协议条款、cookie通知、隐私法和其他与隐私相关的文档中所写的信息的努力，由于缺乏足够的广度和深度来覆盖隐私领域的现有资源而受到影响。作为回应，这个项目正在创建一个基础设施和存储库，用于在线收集大量与隐私相关的文档。这些数据集能够以以前无法实现的覆盖范围和准确性来调查隐私状况，这也支持了法律和公共政策分析。此外，它使研究人员能够建立技术，弥合互联网用户的隐私期望与影响或描述组织隐私实践的文件内容之间的差距。正在讨论的研究课题包括推进法律文本的自然语言处理，识别商业部门的隐私规范和异常值，以及寻找有效地与消费者、研究人员和政策制定者沟通隐私相关文档变化的方法。研究小组正在建立一个大规模的、纵向的、注释的、可搜索的隐私相关文件资源：隐私政策、服务协议条款、cookie通知、美国和世界各地的隐私法、监管指南，以及互联网上的其他相关文本。该团队正在推进自然语言处理（NLP）技术，用于大规模解释与隐私相关的文档，以及以前所未有的规模分析隐私状态，并消除创建工具的障碍，这些工具使用大量关于隐私实践和法规的数据来提供见解和建议。该项目的核心焦点是为隐私研究人员、从业者和政策制定者传播资源，包括搜索引擎、语料库、预训练语言模型、api以及分析工具和结果。这项研究有助于实现长期以来的目标，使消费者、监管机构和其他投资于世界不断发展的信息社会的人能够管理隐私。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

A key aspect of consumer privacy protection is transparency. Yet despite a wealth of privacy policies and other texts that are pertinent to organizations' privacy practices, analyzing those documents has been an obstacle to understanding digital privacy at scale. Efforts to leverage information written in privacy policies, terms of service agreements, cookie notices, privacy laws, and other privacy-related documents suffer from a lack of existing resources with sufficient breadth and depth to cover the privacy landscape. In response, this project is creating an infrastructure and repository for the large-scale collection of privacy-related documents online. The collection enables surveying the privacy landscape with previously untenable coverage and accuracy, which also supports legal and public policy analysis. Additionally, it enables researchers to build technologies that bridge the gap between internet users' privacy expectations and the contents of the documents that influence or describe organizations' privacy practices. Research topics being addressed include advancing natural language processing of legal text, identifying privacy norms and outliers for sectors of commerce, and finding ways to effectively communicate changes in privacy-related documents to consumers, researchers, and policymakers.The research team is building a large-scale, longitudinal, annotated, and searchable resource of privacy-related documents: privacy policies, terms of service agreements, cookie notices, privacy laws in the U.S. and around the world, regulatory guidelines, and other related texts on the internet. The team is advancing natural language processing (NLP) techniques for large-scale interpretation of privacy-related documents, as well as analyzing the state of privacy at an unprecedented scale and removing barriers for creating tools that use large amounts of data about privacy practices and regulations to provide insights and recommendations. A core focus of the project is the dissemination of resources for privacy researchers, practitioners, and policymakers, including search engines, corpora, pre-trained language models, APIs, and analysis tools and results. This research is helping to realize long-standing goals to make privacy manageable for consumers, regulators, and others invested in the world's evolving information society.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.