遇见数据集

Human Influence Telemetry (HIT) v1.0.1 — Poenitentia Institutionum Project

收藏
Zenodo2026-07-11 更新2026-08-02 收录
官方服务:

资源简介:

Human Influence Telemetry (HIT) is a schema-validated framework for diagnosing whether human prudence remains operative inside AI-mediated institutional decisions. HIT operationalizes six canonical dimensions of human deliberative involvement — Counsel, Judgment, Command, Correction, Repair, and Reform — each grounded in verbatim-sourced annotation rows drawn from AI governance case law, regulatory guidance (EU AI Act Article 14, GDPR Article 22, NIST AI RMF), and documented institutional failures (Toeslagenaffaire, UnitedHealth NH, Cigna PxDx). This release (v1.0.1) is a patch update to the Poenitentia Institutionum research repository, the governed source repository supporting a book chapter contribution to AI and the Catholic Perspective: Interventions from the Economic Catholic Front (En Route Books and Media, eds. Prof. Jane Frances Adolphe and Dr. Lukasz Miroslaw). It brings the chapter draft materials and final manuscript in line with the six-dimension HIT schema (Reform) already formalized in v1.0.0 — no new schema, dimension, or conceptual content was introduced. Changes since v1.0.0: Chapter draft (Section III) and Guiding Principles updated from five to six telemetry questions, adding Reform Final manuscript footnote renumbering to accommodate a new footnote describing the Reform dimension and citing the v1.0.0 repository release README.md and CITATION.cff updated with Zenodo DOI badge and citation metadata Poenitentia Institutionum (institutional repentance) is an analogical Catholic governance framework, drawn from the Council of Trent's penitential structure, for institutional repair when human prudential deliberation has been diminished by AI-mediated analysis and evaluation. This archive packages the full v1.0.1 repository state, including the HIT JSON/YAML schema and dimension specification, the fully reconciled HIT annotation set, an interactive HIT dashboard, chapter draft materials, the final manuscript, CI-validated schema tooling, and governance disclosure and source registry records. The repository is governed under the Applied AI Research Translator method (DOI: 10.5281/zenodo.20478121).

提供机构:
Zenodo
创建时间:
2026-07-05
二维码
社区交流群
二维码
科研交流群
商业服务