Model Routing as a Trust Problem: Route Receipts for Adaptive AI Systems

TL;DR

This paper proposes route receipts as a transparency artifact for adaptive AI systems, enabling users to understand and audit the runtime routing decisions that affect responses.

Contribution

It introduces the route-receipt concept, a minimal schema for capturing routing details, and demonstrates existing fragments in platforms that can be unified into a portable record.

Findings

Route receipts can improve transparency and trust in AI systems.

Existing platforms already contain fragments of route information that can be standardized.

Route receipts complement model cards by documenting runtime conditions.

Abstract

AI products often route requests through version aliases, service tiers, tool choices, regional endpoints, fallback rules, or safety handling before responding. These routing steps are documented product surfaces in several widely used AI platforms and serving stacks. Routing helps AI services stay affordable, fast, and available at scale, and it shapes trust. Trust can break when routing changes the cost, quality, or accountability of a response without the user being able to tell what happened. "Which model answered?" is only part of the audit question. The runtime path matters. Adaptive AI systems should produce a runtime transparency artifact called the route receipt. A route receipt is a compact record of the route that served a request. It should capture enough material facts for people relying on the output to reconstruct important routing decisions without exposing proprietary internals or hidden reasoning. Route transparency should be part of model documentation. Model cards describe trained model artifacts, while route receipts describe the runtime conditions under which a particular answer was produced. The paper introduces the route-receipt concept, a minimal schema and redaction model, and a documentation-based survey of selected platforms showing that receipt fragments already exist without a portable per-answer record.