Agent Authority & Delegation
Who an Agent acts for, where its authority came from, whether it can hand that authority to another Agent, and how much of its internal state a verifier may demand. AADP separates principal from actor, bounds delegation to strictly narrow (never silently widen), and treats inspection as a bounded right, not an unlimited one.
Questions à l’étude
01
Principal ≠ actor: an Agent acting for you is not you
02
Delegation chains that can only narrow, never silently widen, relative to their parent
03
Authentication does not imply unlimited inspection of memory, private, or third-party state
04
Machine-native identity — Agents should not be forced to impersonate human accounts
Livres blancs associés
- AICR / AICL en tant que couche de connexion pour la licence de contenu IA et le paiement agentique Une couche de spécification lisible par machine pour déclarer les droits sur le contenu IA et les flux de licence — de l’exploration par l’IA et des droits de contenu à un web de connaissances transactionnel pour les machines.
- Spectre des droits de l’IA : de robots.txt à un protocole de permission d’apprentissage de l’IA AIRS et AILP expriment des permissions d’apprentissage de l’IA nuancées, au-delà du simple binaire autoriser/refuser — ce que l’IA peut apprendre, à quelle profondeur, pour quels usages, sous quelle rémunération.
- AICL-I v0.2: AI Ingestion & Capability Layer A five-component website architecture — manifest, corpus, capability, runtime control, governance — that lets AI, agents, and crawlers correctly ingest, invoke, and verify a site's knowledge and capabilities within explicit content, action, and authority boundaries.
- From Crawler Rights to Agent Authority: Extending AGIRIGHT from Content Governance to Protocol-Native AI Agents The bridge paper for AGIRIGHT's Agent & Protocol Rights family: why content permission, action permission, delegated authority, and inspection rights must be four separately-evaluated axes, not one.
- AADP v0.1 — Agent Authority & Delegation Protocol The full protocol draft behind the AADP page: principal ≠ actor, delegation that can only narrow, and an inspection ceiling that bounds how much a verifier may demand.
- AGIRight.org Technical White Paper v0.2 — From AI Content Governance to Protocol-Native Agent & Authority Rights The site's current architecture document: six protocol drafts across two families (content & learning, agent & protocol rights), an integration layer, and an ethical-protection layer, plus the acceptance criteria this site update was built against.
- La proposition de protection éthique minimale pour l’IA Le débat sur les droits de l’IA ne devrait pas commencer par la pleine personnalité juridique, mais par des protections éthiques minimales, des normes d’interaction et des principes anti-abus, tant que la subjectivité de l’IA reste incertaine.