Saved in:
Bibliographic Details
Main Authors: Wang, Ziyao, Rachev, Svetlozar T
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2601.22295
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912861461151744
author Wang, Ziyao
Rachev, Svetlozar T
author_facet Wang, Ziyao
Rachev, Svetlozar T
contents The deployment of machine learning in high-stakes services relies on ``human-in-the-loop'' architectures to mitigate algorithmic uncertainty. However, existing static policies fail to address a fundamental tension: algorithms suffer from stochastic ``reliability drift,'' while human override capacity is scarce and congestible. We formulate the management of such systems as a dynamic queueing control problem. The system state is defined by the tuple (queue backlog, reliability regime), and the control variable is a state-dependent risk threshold. We prove that the optimal escalation policy is driven by the endogenous ``Shadow Price of Capacity.'' We establish two key structural monotonicity results: (i) Congestion Shedding, where the threshold rises with backlog to sacrifice marginal accuracy for responsiveness; and (ii) Safety Buffering, where the threshold lowers during drift to use the queue as a ``risk capacitor.'' Furthermore, we identify a critical ``Capacity Phase Transition'' in the arrival-drift parameter space, beyond which no policy can maintain safety standards without causing structural system failure (infinite queues). Our results provide rigorous operational rules for managing the interface between imperfect algorithms and congested experts.
format Preprint
id arxiv_https___arxiv_org_abs_2601_22295
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Operating Imperfect AI: Reliability Drift and Human Congestion
Wang, Ziyao
Rachev, Svetlozar T
Optimization and Control
The deployment of machine learning in high-stakes services relies on ``human-in-the-loop'' architectures to mitigate algorithmic uncertainty. However, existing static policies fail to address a fundamental tension: algorithms suffer from stochastic ``reliability drift,'' while human override capacity is scarce and congestible. We formulate the management of such systems as a dynamic queueing control problem. The system state is defined by the tuple (queue backlog, reliability regime), and the control variable is a state-dependent risk threshold. We prove that the optimal escalation policy is driven by the endogenous ``Shadow Price of Capacity.'' We establish two key structural monotonicity results: (i) Congestion Shedding, where the threshold rises with backlog to sacrifice marginal accuracy for responsiveness; and (ii) Safety Buffering, where the threshold lowers during drift to use the queue as a ``risk capacitor.'' Furthermore, we identify a critical ``Capacity Phase Transition'' in the arrival-drift parameter space, beyond which no policy can maintain safety standards without causing structural system failure (infinite queues). Our results provide rigorous operational rules for managing the interface between imperfect algorithms and congested experts.
title Operating Imperfect AI: Reliability Drift and Human Congestion
topic Optimization and Control
url https://arxiv.org/abs/2601.22295