Regulating Preconditions for Algorithmic Harm

Abstract

Abstract Previous chapters identified several preconditions for algorithmic harm. Chapter 8 considers policies that target these preconditions. First and foremost, algorithmic harm is especially prevalent in U markets, where most consumers suffer from information and rationality deficits. It is therefore natural to consider policies that attempt to increase information and to reduce the impact of behavioral biases. Second, AI-powered algorithms need a lot of data to engage in the harmful activities described in previous chapters, such as price discrimination and product targeting. Policies that seek to increase privacy and data security can thus reduce algorithmic harm. Finally, algorithmic harm (for example, price discrimination or enhanced product targeting and differentiation) often requires a certain degree of market power. Accordingly, policies that limit market power, as in the domain of antitrust law, can reduce algorithmic harm.

Ähnliche Arbeiten