Democratic legitimacy in algorithmic governance:

first parameters for its application in the use and regulation of Data and Artificial Intelligence

Abstract

Algorithms are increasingly playing a crucial role in governing various economic, political, and social relationships in modern democracies. Activities such as content moderation on social media platforms, worker management on applications, and the provision of public services are already being affected by what is known as “algorithmic governance”. Often, this type of governance is criticized for violating due process rules, concentrating decision-making power in technology corporations, and producing impacts that prioritize the interests of a social and economic elite over other segments of the population. Given the growing concerns about a potential “dictatorship” of algorithms, it is imperative that this new form of governance, typical of the digital era, be analyzed from the perspective of democratic values. The goal of the article is to demonstrate how democratic legitimacy can be a useful analytical tool for evaluating algorithmic governance practices, as well as guiding the development, regulation, and research of artificial intelligence. The concept of democratic legitimacy (Scharpf, 1999; Schmidt, 2013; Haggart & Keller, 2021) prompts us to analyze the legitimacy of the actors, the processes, and the results of algorithmic governance, leveraging compliance of these practices to parameters of representativeness, due process, and social justice. From this particular analytical perspective, the article highlights initial efforts to democratize algorithmic governance and the need to expand a research agenda in this field.