A Global Comparison of Schemas, Transparency, and Interoperability in Public-Sector AI Registers and Inventories
A study of public-sector AI registers and inventories finds missing data and proposes a layered visibility framework for transparency and interoperability.
This paper compares 8,368 public-sector AI records across 72 countries and finds broad schemas have substantial missingness, with limitations in transparency and interoperability, proposing a layered visibility framework.
- Study of public-sector AI registers and inventories for transparency and interoperability
- 8,368 records across 72 countries compared across 23 harmonized fields
- Identifies missingness in register schemas and need for shared concepts and provenance
- Layered visibility framework for register disclosure arrangements
Full article114 words · extracted from arxiv.org · click to collapse
Artificial intelligence (AI) registers and inventories aim to make governmental AI visible, but their institutional scope, schemas, and reporting practices construct different representations of public-sector AI. We compare 8,368 records from country-specific and transnational inventories covering 72 countries. Across 23 harmonized fields, registers shared a descriptive core but rarely requested information about appeals, risks, legal bases, or external evaluation. We found that broad schemas often contained substantial missingness, schema similarity showed no significant patterned convergence, and multiple sources covering the same jurisdictions overlapped only selectively. Based on these findings, we synthesize a layered visibility framework that shows how register records reflect disclosure arrangements and why interoperability requires shared concepts, clear definitions, and preserved provenance.
Text extracted automatically; images, tables and formatting may be missing. Original: https://arxiv.org/abs/2609.24883