Hadith-Inspired Framework for Claim Provenance in AI Systems
A new paper on arXiv (2607.24117) proposes an operational framework for claim-level provenance in multi-agent knowledge systems, drawing on classical Islamic hadith science. The framework addresses the problem of knowledge transmitted through chains of autonomous transformations, where existing provenance work records execution traces and source-reliability estimation but lacks graded, per-domain transmitter reliability with completeness semantics. The authors adapt the concepts of isnad (complete transmission chain) and rijal (systematic grading of narrators' integrity and precision) to attach graded reliability to claim-level transmission chains, with transformation-typed aggregation, decoupled content criticism, and serve/review/quarantine routing. The paper is categorized as a new submission on arXiv and does not specify authors or institutions.
Key facts
- Paper arXiv:2607.24117 proposes a framework for claim-level provenance in multi-agent knowledge systems.
- Framework is inspired by classical Islamic hadith science, specifically isnad and rijal methodologies.
- Existing provenance work records execution traces, tool calls, and evidence links.
- Source-reliability estimation methods like truth discovery and reputation systems are established.
- The framework includes graded, per-domain transmitter reliability, completeness semantics, and transformation-typed aggregation.
- It features decoupled content criticism and serve/review/quarantine routing.
- The paper is a new submission on arXiv with no listed authors or institutions.
- The approach addresses knowledge accumulation through chains of autonomous transformations.
Entities
Institutions
- arXiv