CAIDP Europe Submits Expert Analysis on AI Social Scoring Prohibition to Dutch Data Protection Authority

The Center for AI and Digital Policy (CAIDP) Europe has submitted comprehensive comments to the Dutch Personal Data Authority on the social scoring prohibition in EU Regulation 2024/1689 (AI Act). The submission addresses key aspects of AI social scoring systems and their impacts on fundamental rights, privacy, and democratic values.

Key Points from CAIDP's Response:

1.     Legal Framework:

·       Analysis of social scoring prohibition under EU AI Act

·       Emphasis on maintaining unambiguous ban without regulatory loopholes

·       Support from international frameworks including UNESCO and Universal Guidelines for AI

2.     Documented Cases of Concern:

·       Dutch childcare benefits scandal affecting low-income families

·       Rotterdam's welfare authorities' discriminatory algorithms

·       French Social Security Agency's CNAF system impacts on marginalized groups

·       Meta's "Special Ad Audience" tool discrimination in housing

·       Austrian AMAS job seeker classification system

·       World Bank's Social Card registry in Serbia

·       UK's Department for Work and Pensions' Universal Credit issues

3.     Critical Concerns:

·       Ambiguity between known, inferred, and predicted characteristics

·       Cross-contextual data usage risks

·       Disproportionate impact on vulnerable populations

·       Need for clear enforcement mechanisms

·       Risks of data repurposing

4.     Key Recommendations:

·       Establishing concrete criteria for assessing social scoring systems

·       Developing independent review mechanisms

·       Creating clear guidelines for data repurposing

·       Strengthening oversight and enforcement processes

·       Addressing loopholes in Article 6 regarding derogation