Skip to main content
TopAIThreats home TOP AI THREATS
Harm Mechanism

Social Scoring

AI systems that assign scores to individuals based on behaviour, social connections, or personal characteristics, used to determine access to services, opportunities, or freedoms.

Definition

Social scoring refers to systems that aggregate data about individuals’ behaviour, social relationships, financial activity, and other attributes to compute a numerical score or rating used for governance, access control, or social stratification. The most widely discussed example is China’s Social Credit System, which links behavioural assessments to consequences including travel restrictions, loan eligibility, and public reputation. However, social scoring practices are not limited to any single jurisdiction. Insurance companies, landlords, employers, and financial institutions in many countries use AI-driven scoring systems that evaluate individuals based on behavioural and social data, often without the transparency or accountability safeguards applied to traditional credit scoring.

How It Relates to AI Threats

Social scoring is a significant concern within the Privacy and Surveillance Threats domain, particularly the mass-surveillance-amplification sub-category. AI enables social scoring systems to process vastly more data sources, identify subtler behavioural patterns, and operate at population scale with minimal human oversight. These systems can incorporate data that individuals did not knowingly provide, draw inferences about character or trustworthiness from indirect signals, and impose consequences that are difficult to understand or contest. The EU AI Act explicitly prohibits social scoring by public authorities when it leads to detrimental treatment unrelated to the context in which data was collected, recognising it as an unacceptable risk to fundamental rights.

Why It Occurs

  • Governments and institutions seek efficient, automated methods for evaluating populations and allocating resources
  • AI enables integration of diverse data sources into unified behavioural assessments at population scale
  • Digital infrastructure generates continuous streams of scoreable data through everyday transactions and interactions
  • Opacity of scoring algorithms prevents individuals from understanding or challenging their assessments
  • Insufficient regulatory frameworks in many jurisdictions allow scoring practices to proliferate without oversight

Real-World Context

China’s Social Credit System remains the most prominent example of state-administered social scoring, affecting millions of citizens’ access to travel, education, and financial services based on algorithmically assessed behaviour. Beyond state systems, private-sector social scoring operates in many countries: tenant screening services score rental applicants, insurance companies adjust premiums based on behavioural data, and platform-based reputation systems affect workers’ access to employment. The EU AI Act’s prohibition on social scoring by public authorities reflects a regulatory judgment that such practices are fundamentally incompatible with democratic values and human dignity.

Last updated: 2026-02-14