AI governance transparency limited by algorithm registers, study finds

A new paper explores the limitations of algorithm registers in providing transparency for public services, particularly in understanding the sociotechnical systems that govern algorithmic decision-making. Through a case study involving a Dutch city's algorithm register and a decision-support tool for welfare benefits, researchers conducted participatory system mapping workshops with various stakeholders. The findings indicate that algorithm registers alone fail to capture crucial safety hazards and governance issues, highlighting the need for broader stakeholder engagement to achieve effective AI governance and accountability. AI

IMPACT Highlights the need for more comprehensive methods beyond simple registers for effective AI governance and accountability in public services.

RANK_REASON Academic paper on AI governance and system transparency. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• Algorithm registers
• arXiv
• Dutch City

• paper
• policy

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →