Answers are grounded in this paper and cite their sources.
AI agents and other non-human identities depend on public-key cryptography for authentication, token signing, certificates, key exchange, software provenance, and secure service-to-service communication. Future quantum computers could undermine the RSA and elliptic-curve mechanisms these depend on, while data with a long confidentiality life is already exposed to harvest-now-decrypt-later collection. This paper proposes a crypto-agile migration architecture that takes the agent identity lifecycle, rather than the application or the network protocol, as the unit of migration. It inventories cryptographic dependencies and links each to agent ownership, privilege, data sensitivity, autonomy, recovery requirement, and revocation path; classifies and prioritizes them by confidentiality lifetime and exposure; establishes crypto-agile control points that externalize algorithm choice from application code; validates post-quantum capability against certificate, token, and message size limits before any cohort moves; runs hybrid or dual-trust operation through a bounded compatibility window; migrates by risk-ordered cohort with defined stop conditions; and retires classical-only trust only once rollback, audit, and recovery criteria are satisfied. The architecture is designed and has not been validated as a completed enterprise migration. Algorithm parameter sizes used in the analysis are taken from the published standards; migration percentages, downtime, and performance impact are not claimed and are identified as pilot outputs.
No takes yet. Share an insight, caveat, or question.
Parijat Sharma (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: