We address the potential of a participatory algorithm design that incorporates stakeholder deliberation and can be used as a technique for “future-proofing”—surfacing past and present human biases so as to improve future decision-making processes, whether algorithmic or human.
Past work exploring algorithmic fairness has focused on how to define fairness and how to create fairer algorithms. These efforts include designing participatory algorithms or auditing algorithms for disparate outcomes. However, participatory algorithm designs have yet to explore integrating stakeholder deliberation, and algorithmic auditing may be limited as a reactive approach.
