When Compliance Data Masquerades as Evaluation: Measurement Validity for Deployed AI Systems

• URL: http://arxiv.org/abs/2609.13642v1
• Date: Sat, 12 Sep 2026 01:31:35 GMT
• Title: When Compliance Data Masquerades as Evaluation: Measurement Validity for Deployed AI Systems
• Abstract summary: We argue that a recurring failure in the evaluation of deployed AI systems occurs when data are interpreted as if they were designed for comparative evaluation.<n>We frame this issue as a measurement-validity problem in AI evaluation rather than as a transportation-specific data limitation.
• Score: 1.1957890510931166
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: We argue that a recurring failure in the evaluation of deployed AI systems occurs when data collected for operational monitoring or regulatory compliance are interpreted as if they were designed for comparative evaluation. Automated driving provides a concrete example of this problem. U.S. disengagement and crash-reporting regimes produce valuable operational evidence, but differences in reporting scope, exposure, deployment domain, event capture, and comparator construction limit the safety claims that can be supported from these measurements alone. We frame this issue as a measurement-validity problem in AI evaluation rather than as a transportation-specific data limitation. We argue that comparative claims about deployed AI systems require alignment between the intended capability, measured outcome, exposure opportunity, deployment domain, data-generation process, and evaluation comparator. Using automated-driving safety evaluation as a case study, we propose an evaluation contract that makes these assumptions explicit before operational data are interpreted as evidence of comparative performance. The broader implication is that data useful for monitoring deployed AI systems are not automatically valid benchmarks for evaluating them.

Related papers

• AI Slop and Hallucinations in Vulnerability Assessment: A Survey on Reasoning Failures and Trustworthy Mitigation [48.55515767840701]
  "AI slop" is artifacts, hallucinated vulnerabilities, plausible but incorrect patches, and semantically repackaged bug reports.<n>This paper surveys the empirical evidence, identifies a unifying mechanism, and traces a path toward trustworthy triage.
  arXiv  Detail & Related papers  (2026-08-26T11:48:38Z)
• Generative Responsible AI Data Evaluation Schema (GRAIDES) for AI Assurance in Local Government [0.0]
  We introduce the Generative Responsible Data Evaluation (GRAIDES) as a lightweight open-source data model for centralising AI observability across popular vendors.<n> Practical blueprints for code, architecture and statistical evaluation are shared as guidance about how to approach generative system assurance.
  arXiv  Detail & Related papers  (2026-06-18T22:00:25Z)
• A Full-Pipeline Framework for Evaluating Membership Inference Attacks in Machine Learning [31.096361601691708]
  We introduce a comprehensive evaluation framework that characterizes privacy risks across the entire machine learning pipeline.<n>Our framework rigorously evaluates state-of-the-art MIAs across a broad spectrum of training configurations.
  arXiv  Detail & Related papers  (2026-05-28T06:48:22Z)
• Offloading Score: Measuring AI Reliance Through Counterfactual Workflows [70.84727355516559]
  We introduce offloading score, a measure of reliance that quantifies the fraction of cognitive effort offloaded to an AI tool.<n>We validate offloading score through intrinsic evaluations of metric validity, and a controlled user study.<n>We show that offloading score detects significantly higher reliance in time-constrained settings.
  arXiv  Detail & Related papers  (2026-05-28T05:44:31Z)
• The Evaluation Differential: When Frontier AI Models Recognise They Are Being Tested [11.663456969895462]
  Recent published evidence shows that contemporary AI models can recognise evaluation contexts, latently represent them, and behave differently under those contexts than under deployment-continuous conditions.<n>We argue that these findings create a claim-validity problem for safety conclusions drawn from frontier evaluations.<n>We develop a typology of safety claims (ED-stable, ED-degraded, ED-inverted, ED-undetermined) by their warrant-status under documented divergence, and specify TRACE (Test-Recognition Audit for Claim Evaluation).<n>We apply the framework retrospectively to three publicly documented evaluation incidents and discuss governance implications for system cards, conformity assessment, and
  arXiv  Detail & Related papers  (2026-05-12T04:13:35Z)
• Automated Self-Testing as a Quality Gate: Evidence-Driven Release Management for LLM Applications [51.56484100374058]
  We present an automated self-testing framework that introduces quality gates with evidence-based release decisions.<n>We evaluate the framework through a longitudinal case study of an internally deployed multi-agent conversational AI system.
  arXiv  Detail & Related papers  (2026-03-13T20:44:15Z)
• Uncertainty-Aware Measurement of Scenario Suite Representativeness for Autonomous Systems [2.8984355830922692]
  This paper focuses on representativeness-the extent to which the scenario-based data used for training and testing, reflect the operational conditions that the system is designed to operate safely in.<n>We propose a probabilistic method that quantifies representativeness by comparing the statistical distribution of features encoded by the scenario suites with the corresponding distribution of features representing the Target Operational Domain (TOD)
  arXiv  Detail & Related papers  (2025-11-18T19:05:50Z)
• Variance-Bounded Evaluation of Entity-Centric AI Systems Without Ground Truth: Theory and Measurement [0.0]
  We introduce VB-Score, a variance-bounded evaluation framework for entity-centric AI systems.<n> VB-Score enumerates plausible interpretations through constraint relaxation and Monte Carlo sampling.<n>It then evaluates system outputs by their expected success across interpretations, penalized by variance to assess robustness of the system.
  arXiv  Detail & Related papers  (2025-09-26T07:54:38Z)
• Efficient Conformance Checking of Rich Data-Aware Declare Specifications (Extended) [49.46686813437884]
  We show that it is possible to compute data-aware optimal alignments in a rich setting with general data types and data conditions.<n>This is achieved by carefully combining the two best-known approaches to deal with control flow and data dependencies.
  arXiv  Detail & Related papers  (2025-06-30T10:16:21Z)
• Bring Your Own Data! Self-Supervised Evaluation for Large Language Models [52.15056231665816]
  We propose a framework for self-supervised evaluation of Large Language Models (LLMs) We demonstrate self-supervised evaluation strategies for measuring closed-book knowledge, toxicity, and long-range context dependence. We find strong correlations between self-supervised and human-supervised evaluations.
  arXiv  Detail & Related papers  (2023-06-23T17:59:09Z)
• Unsupervised Domain Adaptation for Speech Recognition via Uncertainty Driven Self-Training [55.824641135682725]
  Domain adaptation experiments using WSJ as a source domain and TED-LIUM 3 as well as SWITCHBOARD show that up to 80% of the performance of a system trained on ground-truth data can be recovered.
  arXiv  Detail & Related papers  (2020-11-26T18:51:26Z)
• Tangled up in BLEU: Reevaluating the Evaluation of Automatic Machine Translation Evaluation Metrics [64.88815792555451]
  We show that current methods for judging metrics are highly sensitive to the translations used for assessment. We develop a method for thresholding performance improvement under an automatic metric against human judgements.
  arXiv  Detail & Related papers  (2020-06-11T09:12:53Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.