Dataset-Driven Policy: An Open Repository Linking AI Deployments, Regulatory Interventions, and Outcomes

Despite calls for evidence-based policy (see Booth, 2025; Gupta et al., 2024), there’s a lack of comprehensive datasets that trace the lifecycle from AI deployment through regulatory action to real-world impacts (good or bad). This research would assemble, standardize, and publish such a dataset, drawing on public records, agency reports, media analyses, and expert interviews. For each case, it would document: the AI system and its context; the oversight/intervention applied (e.g., audit, recall, policy update); and the outcomes (e.g., reduction in bias, changes in patient outcomes, public trust metrics). This would allow for meta-analyses of what kinds of oversight “work” in practice, and under what conditions, addressing the gap highlighted by Raji (2022) and Adepoju & Chinonyerem (2025). By making the data open, this project would catalyze a new wave of empirical, comparative research on AI governance effectiveness.

References:

1. From Algorithmic Audits to Actual Accountability: Overcoming Practical Roadblocks on the Path to Meaningful Audit Interventions for AI Governance. Inioluwa Deborah Raji (2022). AAAI/ACM Conference on AI, Ethics, and Society.
2. ADVANCING GOOD GOVERNANCE THROUGH AI-POWERED OVERSIGHT IN THE UNITED STATES: RISKS AND OPPORTUNITIES FOR PUBLIC INSTITUTIONS. ADEWUMI SUNDAY ADEPOJU, CONFIDENCE ADIMCHI CHINONYEREM (2025). Journal of African Advancement and Sustainability Studies.
3. AI Governance and Lessons Learned as an AI Policy Advisor in the United States Senate. Serena Booth (2025). AAAI Conference on Artificial Intelligence.
4. Data-Centric AI Governance: Addressing the Limitations of Model-Focused Policies. Ritwik Gupta, Leah Walker, Rodolfo Corona, Stephanie Fu, Suzanne Petryk, Janet Napolitano, Trevor Darrell, A. Reddie (2024). arXiv.org.