Uncovering Incentives for Implementing AI Governance Programs: Evidence From the Field

Methods and applications of artificial intelligence (AI) are increasingly impacting society. To manage this technology's positive and negative effects, stakeholders (governments, firms, and nonprofits) are developing programs such as principles, guidelines, codes, certifications, and standards, among others, as an alternative to regulation or hard law. The literature classifies these programs as soft law, defined as any initiative that creates substantive expectations that are not directly enforced by government. Although these programs are incredibly flexible and can adapt to the constantly changing nature of emerging technologies, their main weakness is their voluntary nature. In other words, organizations are responsible for deciding if the implementation of soft law is in their best interests. The objective of this article is to inform stakeholders of the three incentives that motivate the enforcement or implementation of soft law and illustrate them with examples taken from an analysis of over 600 AI-centric programs.