Who Decides What Online and Beyond: Freedom of Choice in Predictive Machine-Learning Algorithms

The controversial impact raised by the pervasive deployment of online systems based on predictive artificial intelligence (AI) techniques, and specifically machine-learning algorithms (MLA), on human decision-making and (supposedly) free choices is extensively acknowledged as a core topic in the debate and literature in AI and ethics. However, thus far, such impact has been mainly framed in terms of human choice manipulation and privacy infringements, while the ethical issue of human freedom of choice in predictive MLA has been poorly explored per se. This chapter aims to fill this gap and provide a systematic ethical inquiry into the impact raised by predictive MLA on the preconditions underlying the exercise of freedom of choice, online and beyond. To this aim, we first pursue an ethical inquiry into the preconditions securing at a minimum threshold human freedom of choice, drawing insights on main theories developed both in moral philosophy and socio-political philosophy. Then, we analyze how MLA are governing and reshaping the contexts in which we both prepare and make our choices. We show how by doing so MLA negatively affect the preconditions of our freedom of choice and the risks such impact raises in the short and long term, at the individual and collective level. Finally, we address such risks and highlight how our ethical inquiry helps pinpointing some key ethical criteria that ought to be considered and operationalized to secure our freedom of choice as a fundamental ethical-normative value in our contemporary MLA-driven societies.