TRACE: A Metrologically-Grounded Engineering Framework for Trustworthy Agentic AI Systems in Operationally Critical Domains

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autore principale: Zabolotnii, Serhii
Natura: Preprint
Pubblicazione: 2026
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866911648986431488
author Zabolotnii, Serhii
author_facet Zabolotnii, Serhii
contents We introduce TRACE, a cross-domain engineering framework for trustworthy agentic AI in operationally critical domains. TRACE combines a four-layer reference architecture with an explicit classical-ML vs. LLM-validator split (L2a/L2b), a stateful orchestration-and-escalation policy (L3), and bounded human supervision (L4); a metrologically grounded trust-metric suite mapped to GUM/VIM/ISO 17025; and a Model-Parsimony principle quantified by the Computational Parsimony Ratio (CPR). Three instantiations--clinical decision support, industrial multi-domain operations, and a judicial AI assistant--transfer the samearchitecture and metrics across principally different governance contexts. The L2a/L2b separation makes the use of large language models a deliberate design decision rather than an architectural default, with parsimony quantified through CPR. TRACE introduces CPR as a first-class design principle in trustworthy-AI engineering.
format Preprint
id arxiv_https___arxiv_org_abs_2605_03838
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle TRACE: A Metrologically-Grounded Engineering Framework for Trustworthy Agentic AI Systems in Operationally Critical Domains
Zabolotnii, Serhii
Computation and Language
Artificial Intelligence
Human-Computer Interaction
We introduce TRACE, a cross-domain engineering framework for trustworthy agentic AI in operationally critical domains. TRACE combines a four-layer reference architecture with an explicit classical-ML vs. LLM-validator split (L2a/L2b), a stateful orchestration-and-escalation policy (L3), and bounded human supervision (L4); a metrologically grounded trust-metric suite mapped to GUM/VIM/ISO 17025; and a Model-Parsimony principle quantified by the Computational Parsimony Ratio (CPR). Three instantiations--clinical decision support, industrial multi-domain operations, and a judicial AI assistant--transfer the samearchitecture and metrics across principally different governance contexts. The L2a/L2b separation makes the use of large language models a deliberate design decision rather than an architectural default, with parsimony quantified through CPR. TRACE introduces CPR as a first-class design principle in trustworthy-AI engineering.
title TRACE: A Metrologically-Grounded Engineering Framework for Trustworthy Agentic AI Systems in Operationally Critical Domains
topic Computation and Language
Artificial Intelligence
Human-Computer Interaction
url https://arxiv.org/abs/2605.03838