MCA: Transparent and Accountable Decision Systems

自动化和算法决策系统被广泛用于制定公共政策。尽管这些系统在公共领域内被用作自动决策和资源分配工具，但它们往往是不透明的，也没有进行公开审计。可以提供一定程度的理由和责任的细节往往隐藏在与第三方供应商的专有合同中。没有透明和可解释的过程，就不可能核实结果是否满足道德和公平约束。如果没有问责，就不可能对这些制度产生的决策产生信任。这个项目的重点是在公共政策的自动决策系统的背景下的问责制、透明度和可解释性。这项职业中期促进奖(MCA)支持阿梅莉·玛丽安与带来不同专业知识的合作伙伴/导师合作。PI利用她在设计排名系统和决策算法方面的背景，与合作伙伴的法律学者、政策研究人员和社会科学家团队合作，开发决策系统中的透明度和问责制的指导方针和技术。该项目从计算、法律和伦理的角度解决算法决策的问责制。透明度和问责制是建立对自动决策结果的信任并保证公正和公平结果的关键先决条件。这项研究涉及对用于公共政策决策的数据进行二次分析，这些领域包括学校招生、疫苗和器官捐赠名单优先排序、工作申请、公共住房分配和排名选择投票。PI还利用MCA奖深化和创建与政策制定者的新合作，以及与该项目的导师合作伙伴、康奈尔理工大学数字生活倡议小组主任海伦·尼森鲍姆建立新的多学科合作。通过广泛的利益相关者和学者的参与，这项研究更好地理解了什么是公平、透明和负责任的决策，从而增加了对决策过程的信任，并利用算法决策机制造福社会。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

Automated and algorithmic decision systems are used extensively to enact public policy. Despite their use within the public domain as automated decision-making and resource distribution tools, these systems are often opaque and not publicly audited. Details that could provide a measure of reason and accountability are often hidden within proprietary contracts with third-party vendors. Without transparent and explainable processes, it is not possible to verify whether the outcomes satisfy ethical and fairness constraints. Without accountability, there can be no trust in the decisions produced by these systems. This project focuses on accountability, transparency, and explainability in the context of automated decision systems for public policy. This Mid-Career Advancement (MCA)Award supports Amelie Marian to work with a partner/mentor who brings different expertise. The PI leverages her background in designing ranking systems and decision-making algorithms to collaborate with the partner’s team of legal scholars, policy researchers, and social scientists to develop guidelines and techniques for transparency and accountability in decision systems.This project addresses the accountability of algorithmic decisions through the perspectives of computation, law, and ethics. Transparency and accountability are critical prerequisites for building trust in the results of automated decisions and guaranteeing fair and equitable outcomes. The research involves secondary analysis of data used for public policy decisions in domains such as school admissions, vaccine and organ donation lists prioritization, job applications, public housing allocations, and ranked choice voting. The PI is using the MCA award also to deepen and create new collaborations with policy-makers, as well as establish new multidisciplinary collaborations with the project's mentor partner, Helen Nissenbaum, director of the Digital Life Initiative group at Cornell Tech. By involving a wide array of stakeholders and scholars, this research obtains a better understanding of what constitutes a fair, transparent, and accountable decisions, thereby increasing trust in the decision processes and leveraging algorithmic decision mechanisms for social good.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.