Aegon: Auditable AI Content Access with Ledger-Bound Tokens and Hardware-Attested Mobile Receipts
Amrish Baskaran and Nirbhay Pherwani and Raghul Krishnan · 2026
Read Aegon: Auditable AI Content Access with Ledger-Bound Tokens and Hardware-Attested Mobile Receipts by Amrish Baskaran and Nirbhay Pherwani and Raghul Krishnan (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-06693.txt.
Introduction
Baskaran, Pherwani, and Krishnan design Aegon, an audit layer for AI content licensing. Current standards tell AI systems the license terms, but none keep tamper-evident records proving those records were never rewritten. Aegon extends standard JWT tokens with content-specific licensing claims and logs transactions in a Merkle tree over an append-only ledger, so outside auditors can verify entries independently. Publishers check tokens at the edge without a broker, provenance events track content through AI transformation stages, and Android agents add hardware-attested receipts from StrongBox secure elements.
Catalog record
| CREATOR | Amrish Baskaran and Nirbhay Pherwani and Raghul Krishnan |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 31764 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.06693
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Aegon: Auditable AI Content Access with Ledger-Bound Tokens and Hardware-Attested Mobile Receipts Amrish Baskaran and Nirbhay Pherwani and Raghul Krishnan Abstract . Recent standards such as RSL ( 38 ) address AI content policy declaration —telling AI systems what the licensing terms are. However, no existing system provides audit infrastructure —tamper-evident licensing transaction records with independently verifiable proofs that those records have not been retroactively modified. We describe Aegon, a protocol that extends standard JWT tokens with content-specific licensing claims and maintains a Certificate Transparency-style ( 26 ) Merkle tree over an append-only transaction ledger, enabling third-party auditors to independently verify that specific content licensing transactions were recorded and have not been retroactively modified. Publishers validate tokens at the edge using standard JWKS with no broker dependency in the content delivery path. A signed provenance event log tracks content through AI transformation stages (chunking, embedding, retrieval, citation), bound to ledger entries by transaction ID. We further describe hardware-attested compliance receipts for on-device Android AI agents using StrongBox secure element attestation ( 2 ) —to our knowledge, the first application of hardware-attested compliance receipts to AI content licensing. Existing DRM systems use hardware-backed keys for content decryption but do not produce verifiable compliance receipts for audit trails. We describe a reference architecture and an evaluation methodology for measuring protocol overhead. The protocol runs entirely over standard HTTPS and is designed to complement existing licensing standards rather than replace them. 1. Introduction 1.1. Motivation AI systems—large language models used for training, inference, and retrieval-augmented generation (RAG)—depend heavily on web content. Yet no standard mechanism exists for AI platforms to access content with verifiable licensing, and no standard mechanism exists for publishers to enforce or monetize access rights. The result is a coordination failure: • Publishers block all AI traffic or allow unchecked access with no compensation • AI platforms scrape without licenses, facing ongoing lawsuits with billions in potential exposure ( 30 ; 14 ; 7 ) • Bespoke licensing negotiations take months and do not scale ( 29 ) • No audit trail exists to prove content was accessed legally The same problem exists at a second layer: on-device AI agents (Gemini Nano ( 15 ) , Llama ( 27 ) ) access web content during inference on mobile devices with no licensing infrastructure whatsoever. 1.2. Contributions Aegon builds on standard JWT ( 23 ) and JWKS ( 24 ) infrastructure and makes contributions in two areas that, to our knowledge, have no prior art in the content licensing domain: Contribution 1—Auditable Web Protocol Layer: (1) Content-specific licensing claims extending standard JWT—the claim semantics are the contribution, not the validation mechanism. (2) A Certificate Transparency-style ( 26 ) Merkle tree over an append-only transaction ledger, enabling third-party audit verification via Signed Tree Heads and inclusion proofs—to our knowledge, the first application of the CT audit model to content licensing. (3) A signed provenance event log tracking content through AI-specific transformation stages, bound to ledger entries by transaction ID. Contribution 2—Android Attestation Layer: (1) Hardware-attested compliance receipts for on-device AI agents using Android StrongBox secure element ( 2 ) —to our knowledge, the first application of hardware-attested compliance receipts to AI content licensing on mobile devices. (2) Offline proof batching with replay resistance for intermittent mobile connectivity. (3) Broker-side verification of Key Attestation ( 3 ) certificate chains rooted at Google hardware attestation CA. Existing licensing standards (RSL ( 38 ) , AIBDP ( 21 ) ) address policy declaration. Content access protocols (Peek-Then-Pay ( 12 ) , PEAC ( 32 ) ) address the licensing transaction flow. Aegon’s distinct contribution is the audit infrastructure —the Merkle-committed ledger, the cryptographically signed provenance chain, and the hardware-attested mobile receipts—that enables independent verification of licensing transactions and tamper-evident provenance records after the fact. 1.3. Analogous Infrastructure Aegon’s position in the AI content stack is analogous to prior HTTP-based coordination infrastructure: OAuth 2.0 ( 20 ) standardized delegated authentication; Certificate Transparency ( 26 ) introduced Merkle tree audit logs for TLS certificates; PCI DSS established a compliance standard that companies adopt because they need to demonstrate compliance. Aegon adopts CT’s Merkle tree audit construction and Signed Tree Head pattern as a CT-inspired audit design; it does not claim to replicate the full CT ecosystem, and the broker remains a trusted party rather than a trustless log. 2. Background and Related Work The AI content licensing space has seen rapid development since 2024. We organize related work into six categories. 2.1. Emerging Licensing Standards RSL 1.0 ( 38 ) : Published December 2025 by the co-creators of RSS, RSL defines a machine-readable XML vocabulary for content licensing, a Crawler Authorization Protocol (CAP) using license tokens in HTTP headers, and an Open License Protocol (OLP) extending OAuth 2.0. Endorsed by 1,500+ organizations including Cloudflare and Akamai. RSL addresses policy declaration ; Aegon addresses auditability of licensing transactions and tamper-evident provenance evidence . RSL has no audit ledger, no Merkle proofs, no provenance tracking, and no hardware attestation. AIBDP ( 21 ) : An IETF Internet-Draft (August 2025) defining machine-readable AI permissions via HTML meta tags, DNS TXT records, and HTTP headers. A declaration standard with no enforcement mechanism. Agentic JWT ( 18 ) : An IETF Internet-Draft (December 2025) extending OAuth 2.0 for AI agents with cryptographic agent identity. Addresses who the agent is ; Aegon addresses what content the agent can access . Complementary. 2.2. Content Access Protocols Peek-Then-Pay ( 12 ) : An HTTP 203-based protocol with JWT licenses and edge validation. The most architecturally similar system to Aegon. Key difference: Peek-Then-Pay optimizes the discovery and pricing flow; Aegon provides the audit and licensing evidence infrastructure. PEAC ( 32 ) : A verifiable receipt system with compliance proofs and payment integration. PEAC’s receipts overlap with Aegon’s audit trail, but PEAC lacks Merkle tree verifiability and hardware attestation. OpenAttribution ( 31 ) : An open-source framework with DID-based agent identity and session-based content event tracking through AI pipeline stages. Conceptually similar to Aegon’s provenance log, but without cryptographic binding to Merkle-committed ledger entries. TollBit ( 36 ) : A commercial bot paywall and licensing platform ($30.9M total funding including $24M Series A, 500+ publisher sites). Closed-source SaaS without a public protocol specification or audit ledger. Cashmere ( 9 ) : A commercial data infrastructure platform ($5M seed, January 2026) with live deployments at Wiley and Harvard Business Publishing, and Perplexity as both investor and customer. Provides token-level access control, rights management, usage tracking, and revenue settlement. Closed-source with no open protocol specification, no Merkle audit ledger, and no hardware attestation. Its commercial traction validates demand for the infrastructure Aegon describes. 2.3. Traditional Access Control and DRM robots.txt ( 25 ) : Crawler exclusion with no authentication or licensing. OAuth 2.0 ( 20 ; 35 ) : The standard for delegated authentication. Aegon inherits JWKS-based validation ( 24 ) . Aegon’s contribution is the content licensing claim semantics and audit ledger binding. Traditional DRM (Widevine ( 17 ) , FairPlay, PlayReady) protects media at the delivery layer. Aegon differs: (1) no content encryption—audit trails, not access prevention; (2) AI data pipelines, not media playback; (3) JWKS verification, not license server roundtrip. 2.4. Content Provenance C2PA ( 8