Regulatory Alternatives to the AI Social Scoring Ban: A Comparative Perspective

Article 5(1)(c) of the EU AI Act is widely reported as a ban on using AI systems to score individuals’ social behavior, and such ban was likely motivated, at least in some part, by European lawmakers’ concerns over the widely misrepresented dystopian scenario of China’s Social Credit System. However, as the text of the provision presents considerable interpretive uncertainties, it could potentially lead to the prohibition of scoring practices that have been commonplace and socially beneficial both before and after the advent of AI and algorithms. Alternatively, the provision can be interpreted and implemented narrowly—and perhaps must be—so that most existing and potentially implementable scoring practices are not unnecessarily affected. The European Commission’s guidelines publicized in early 2025 suggest that authorities may be leaning towards such narrow reading. A narrow reading is also necessary when one understands the role that social scoring plays in human societies. And by drawing on China’s experiences with implementing and regulating its Social Credit System (SCS), policymakers around the world may also realize that there are feasible options for regulating AI-based social scoring practices instead of insisting on an outright ban.