Collaborative Research: SaTC: CORE: Medium: Methods and Tools for Effective, Auditable, and Interpretable Online Ad Transparency

有针对性的在线广告无处不在。搜索引擎和在线社交网络为广告商提供了强大的目标定位技术，以将其消息传递给特定的最终用户。这些技术帮助用户看到更相关的广告，但也引发了隐私问题。定向广告以各种方式影响人们的生活，例如他们可能遇到的就业，住房和信贷机会。平台最近开始向研究人员和最终用户提供更多关于广告的数据。尽管作出了这些努力，但初步研究发现，目前的透明度机制存在重大缺陷，需要改进方法。该项目旨在更好地了解当前的透明度机制，开发收集透明度数据的新方法，并设计新的机制，使透明度对研究人员、记者、公民社会团体和最终用户更有用，最终提高最终用户对定向广告的信任。首先，研究人员正在绘制广告透明度的当前空间，开发透明度机制的分类，并进行用户研究，以评估这些机制对最终用户和第三方审计员的效用。其次，目前各种社交媒体平台的透明度实施方式非常不同，所提供的数据往往不直接适用于大规模实证研究。研究人员正在开发一种方法，在平台的透明度数据源之上叠加额外的数据，以扩展透明度并标准化跨平台的数据。第三，研究人员正在应用以用户为中心的设计实践来开发和评估一套新的广告透明度机制和用户界面，这些机制和界面是为多个利益相关者设计的：最终用户、公民社会团体和记者。该项目将培养学生在安全、数据科学和人机交互方面的专业知识，这些领域具有广泛的国家重要性。该奖项反映了NSF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。

Targeted online advertising is ubiquitous. Search engines and online social networks provide powerful targeting technologies for advertisers to deliver their messages to specific end users. These technologies help users see more relevant ads, yet raise privacy concerns. Targeted ads impact people's lives in a variety of ways, such as what employment, housing, and credit opportunities they may encounter. Platforms have recently started making transparent to researchers and end-users more data about advertisements. Despite these efforts, initial studies have found significant shortcomings in current transparency mechanisms, necessitating improved methodologies. This project seeks to better understand current transparency mechanisms, develop new methods for collecting transparency data, and design new mechanisms that make transparency more useful to researchers, journalists, civil-society groups, and end-users, ultimately increasing end users' trust in targeted advertising.The research has three key tasks. First, the researchers are mapping the current space of advertising transparency, developing a taxonomy of transparency mechanisms, and conducting user studies to evaluate these mechanisms' utility both for end-users and third-party auditors. Second, transparency is currently implemented very differently by various social media platforms, and the data made available are often not directly suitable for large-scale empirical research. The researchers are developing methodologies to layer additional data on top of platforms' sources of transparency data to extend transparency and standardize data across platforms. Third, the researchers are applying user-centered design practices to develop and evaluate a suite of new ad transparency mechanisms and user interfaces designed for multiple stakeholders: end-users, civil-society groups, and journalists. This project will train students with expertise in security, data science, and human-computer interaction, areas of broad national importance.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.