View and download papers and working papers related to the PRISM standard, all in one place, directly as PDF. All outputs are distributed under CC BY 4.0 (documents) and Apache-2.0 / MIT (code).
A conceptual starting point that defines AI Integrity not as a single score but as a distribution across three axes — value, evidence, and source. It provides the theoretical foundation for turning safety and alignment discourse into measurable infrastructure.
An empirical study that quantitatively validates whether the PRISM vocabulary and code format work on real AI responses across diverse domains.
Defines the format, issuance process, and validation rules of a standard card (the PRISM Risk Signal Card) with which companies and institutions can issue, on a single page, their AI's distribution of values, evidence, and sources along with its blind spots.
How to map and satisfy the logging and transparency requirements of major AI regulatory frameworks — the EU AI Act, NIST AI RMF, ISO/IEC 42001, and Korea's AI Framework Act — using the PRISM standard.
A governance argument for using PRISM's measurable public vocabulary to address a structural flaw in democratic accountability — the problem that AI's political use is expanding while its value hierarchies remain opaque.
A standard specification document defining PRISM's V/E/S/C code format, BNF grammar, vocabulary (Schwartz 19 + Walton 10 + Hovland-Kelley 10), and validation rules.
The first public benchmark to measure and compare V/E/S distributions across 8 frontier models and 366,120 responses. It quantifies the degree of disagreement in value hierarchies between models and automatically detects blind spots by domain and demographic.
A methodology that applies the PRISM standard to embodied AI systems performing physical actions (robots, autonomous driving, medical devices) to enable decision-audit trails.
Proposes a three-tier measurement vocabulary that lets four groups — ethics, clinical, patient, and policy — discuss accountability using the same vocabulary. It argues for PRISM's legitimacy at the governance level.