Quantum-Secure-By-Construction (QSC): A Paradigm Shift For Post-Quantum Agentic Intelligence

Arit Kumar Bishwas; Mousumi Sen; Albert Nieto-Morales; Joel Jacob Varghese

cs.AIMar 12, 2026

Quantum-Secure-By-Construction (QSC): A Paradigm Shift For Post-Quantum Agentic Intelligence

Authors: Arit Kumar Bishwas, Mousumi Sen, Albert Nieto-Morales, Joel Jacob Varghese

Organizations: PricewaterhouseCoopers USA, San Francisco, CA · Independent Researcher, USA, Fremont, CA

Abstract

As agentic artificial intelligence systems scale across globally distributed and long lived infrastructures, secure and policy compliant communication becomes a fundamental systems challenge. This challenge grows more serious in the quantum era, where the cryptographic assumptions built into today's AI deployments may not remain valid over their operational lifetime. Here, we introduce quantum secure by construction, or QSC, as a design paradigm that treats quantum secure communication as a core architectural property of agentic AI systems rather than an upgrade added later. We realize QSC through a runtime adaptive security model that combines post quantum cryptography, quantum random number generation, and quantum key distribution to secure interactions among autonomous agents operating across heterogeneous cloud, edge, and inter organizational environments. The approach is cryptographically pluggable and guided by policy, allowing the system to adjust its security posture according to infrastructure availability, regulatory constraints, and performance needs. QSC contributes a governance aware orchestration layer that selects and combines link specific cryptographic protections across the full agent lifecycle, including session bootstrap, inter agent coordination, tool invocation, and memory access. Through system level analysis and empirical evaluation, we examine the trade offs between classical and quantum secure mechanisms and show that QSC can reduce the operational complexity and cost of introducing quantum security into deployed agentic AI systems. These results position QSC as a foundational paradigm for post quantum agentic intelligence and establish a principled pathway for designing globally interoperable, resilient, and future ready intelligent systems.

Figures & tables

Figure 1: Quantum-secure-by-construction agentic AI system model.

Region / Jurisdiction	Layer 1: PQC (Kyber, Dilithium)	Layer 2: QRNG	Layer 3: QKD	Compliance Notes
United States (FIPS/NIST [ 52 ] )	Full support (NIST standard)	Supported in national clouds and DoD networks	Restricted to defense/research use	FIPS-140 compliance required for operational systems
European Union (ENISA/GDPR) [ 53 ]	Fully supported	Available in some member states	Limited by cross-border data sovereignty laws	GDPR & eIDAS integration; ENISA PQC draft alignment
China (Commercial Encryption Law) [ 54 ]	Partial (non-NIST PQC under review)	National QRNG standards vary	Strong QKD support (Beijing $\leftrightarrow$ Shanghai)	Local encryption vendors must be certified
Japan	Fully supported	R&D deployment in Tokyo cloud centers	Emerging satellite-QKD projects	METI and MIC jointly defining crypto standards [ 55 ] [ 56 ]
Australia	Fully supported	Limited QRNG deployment	Under review for GovCloud and critical infrastructure	ASD-approved PQC needed for national defense AI [ 57 ]
Hybrid / Multinational Deployments	Mandatory baseline for all agents	Conditional per-node availability	Optional—negotiated QKD tunnels per jurisdiction	Crypto-pluggable fallback to PQC-only mode

Table 1: Interoperability regimes across jurisdictions and quantum security layers

Metric	PQC Mean	Classical Mean	Unit	Overhead (%)
Encapsulation/Exchange	0.082	0.018	ms	78.2
Decapsulation	0.014	N/A	ms	N/A
Key Generation	0.11	0.01	ms	710.0
Message Signing	0.34	0.01	ms	2385.1
Signature Verification	0.08	0.03	ms	149.6
KEM Public Key Size	1184	65	bytes	1721.5

Table 2: Comparative Performance and Size of PQC vs. Classical Cryptography. Metrics are grouped by key exchange performance, digital signature timing, and cryptographic artifact sizes.

Layer	Parameter	Value	Unit
QRNG (Layer 2) Parameters - Measured
	Nonce Generation Time	1.157	ms
	Observed Throughput	$\approx 28$	Mbps
QKD (Layer 3, BB84 Protocol) Parameters - Simulated
	Encrypt/Decrypt Time (AES)	0.005	ms
	Key Establishment Time	10.0	ms

Table 3: Measured Hardware Latencies and Derivation Parameters

ID	Communication Flow	PQC	QRNG	QKD	Time	Overhead
					(ms)	(%)
1	User $\rightarrow$ Orch. (Request)	Y	Y	N	1.06	6.6
2	Orch. $\leftrightarrow$ Agent (Handshake)	Y	N	N	0.41	693.5
3	Orch. $\rightarrow$ MCP (Task Publish)	Y	Y	Y	2.26	12.7
4	Agent $\leftarrow$ MCP (Fetch Task)	Y	N	Y	1.10	8.8
5	Agent $\rightarrow$ MCP (Result)	Y	Y	Y	3.38	68.8

Table 4: Performance Benchmark of Seven Critical Quantum-Secured Communication Channels

Figure 2: Absolute latency comparison of quantum-secured vs. classical protocols (ms).

ID	Communication Channel	Security Stack	Mean Cloud Latency (ms)
1	User $\rightarrow$ Orch. (Request)	PQC-TLS	326.00
2	Orch. $\leftrightarrow$ Agent (Handshake)	PQC-TLS	297.41
3	Orch. $\rightarrow$ MCP (Task Publish)	PQC-TLS + QKD	299.26
4	Agent $\leftarrow$ MCP (Fetch Task)	PQC-TLS + QKD	298.10
5	Agent $\rightarrow$ MCP (Result)	PQC-TLS + QKD	300.38
6	Agent $\leftrightarrow$ Agent (Memory)	PQC-TLS + QKD	301.53

Table 5: Cloud-Scale Performance: Azure Deployment Latency (PQC-TLS 1.3 with Modeled QKD)

Figure 3: Scalability and throughput of the quantum-secured framework.

Operational Step	Architecture Section(s)	Agent Subsystem / Orchestrator Role	Quantum Security Controls
Quantum Secured Agentic AI System	All Layers (Global Coordination Layer, Agent Layer, Quantum Cryptographic Stack)	All orchestrators and agent subsystems across the pipeline	Full-stack PQC (Kyber, Dilithium), QRNG-sampled state transitions, QKD (if available)
Quantum Secured Session Bootstrap	Global Coordination Layer Quantum Cryptographic Stack	Orchestrator: Session Manager, Crypto Negotiator	Session key exchange (Kyber), agent identity signing (Dilithium), QRNG token, optional QKD
Initialize Session	Global Coordination Layer Quantum Cryptographic Stack	Orchestrator: Identity Enforcer, Key Registrar	PQC-based key provisioning, QRNG-derived session ID, quantum entropy seeding
Secure Client Request	Global Coordination Layer Perception & Input Interface	Orchestrator: Input Validator Agent: Input Filter Module	PQC TLS, Dilithium client signature check, QRNG request hash, PII detection and redaction
Build Task Graph	Global Coordination Layer Planning and Reasoning Engine Securing MCP-Agent Communication	Orchestrator: Task Planner, DAG Compiler Agent: Orchestration Router	QRNG for DAG node IDs, signed edge policies via Dilithium, Kyber-secured task routing
Secure Agent Execution	Agent Layer: RAG Module Memory Store Tools Interface MCP Channel	Agent: Tool Invoker, Retriever, Memory Engine	Agent-to-agent communication via PQC (Kyber), per-agent QRNG token, encrypted tool calls

Table B.1: Mapping of Algorithm Steps to Architecture Sections, Roles, and Quantum Security Controls

#	Domain	Use Case	Agent Roles	PQC Usage	QRNG Usage	QKD Usage
1	Finance & Banking	Decentralized Financial Infrastructure	Trading bots, Compliance agents, Arbitrage agents	Digital signatures, Transaction integrity	Pricing tokens, Trade nonces	Central bank $\leftrightarrow$ Exchange security
2	Aerospace & Space	Space-Earth Autonomy	Telemetry processors, Navigation agents, Signal relayers	Signed commands, Fault tolerance protocols	Command entropy, Data integrity tokens	Earth $\leftrightarrow$ satellite laser-based QKD
3	Healthcare & Life Sciences	Global Health AI	Federated trainers, Diagnosis support agents	Model update encryption, GDPR/HIPAA compliance	Federated round key generation	Optional: Hospital-to-hospital secure sync
4	Energy & Utilities	Smart Grids & Autonomous Utilities	Relay controllers, Load balancers, Demand forecasters	Telemetry protection, Load switch verification	Spoof prevention, Power dispatch randomness	Cross-grid or regional coordination QKD
5	Cybersecurity & Intelligence	Intelligence & Threat Hunting	Threat modelers, SOC analysts, Packet inspectors	Threat signature validation, Data sealing	Forensic log entropy, Attack trace prevention	NATO or inter-agency classified coordination
6	Transportation & Logistics	Secure Global Supply Chains	Asset trackers, Route planners, Customs clearance agents	Shipment signing, Smart contract enforcement	Unique tag tokens, Anti-counterfeiting	Port authority $\leftrightarrow$ regulatory coordination

Table B.2: Quantum-secured multi-agent use cases across domains.

Security Dimension	Classical Crypto	Post-Quantum Crypto (PQC)	QKD + QRNG (Quantum-Resilient)
Resistance to Quantum Attacks	Low – Vulnerable	High – based on lattice, hash, or code hardness [ 70 ] [ 71 ]	Highest – Information-theoretic security; passive detection of eavesdropping [ 72 ]
Forward Secrecy Longevity	Moderate	Strong (at PQC level)	Superior – symmetric keys refreshed via QKD
Entropy Assurance	Weak – PRNG risk	Better, but software-limited	Physical-grade entropy via QRNG, near-ideal unpredictability
Key Distribution Infrastructure	Universal (IP/TLS)	Fully compatible (TLS-PQC-ready) [ 73 ] [ 70 ]	Requires dedicated hardware (fiber/satellite), limited reach
Vulnerability to Eavesdrop/Replay	High	Low	Very Low – quantum disturbances detectable
Implementation Complexity	Low	Moderate	Very High – complex hardware and calibration requirements

Table B.3: Comparison of Security Dimensions for Classical, PQC, and QKD+QRNG Systems

Figure C.1: Logical System Architecture: A Zero-Trust, quantum-secured multi-agent system enforcing cryptographic boundaries across seven critical communication channels.

Figure C.2: Azure PQC Sidecar Architecture: Securing inter-agent communication via transparent PQC-TLS tunnels.

System Specification	Value
Processor	Apple M3 Pro, 3.8 GHz, 12 cores
Memory (RAM)	36 GB
Operating System	macOS 15.2

Table C.1: Computing Environment Specifications

Specification	Value
Random Bit Rate	4 Mbit/s $\pm$ 10%
Thermal Noise Contribution	$<1\%$
Storage Temperature	-25 ∘ C to +85 ∘ C
USB Specification	2.0
Power	Via USB port

Table C.2: ID Quantique Quantis USB 4M QRNG Specifications

Attack Vector	Injected	Detected/Blocked	Efficacy
Integrity Violation (Tamper)	50,000	50,000	100.0%
Replay/Freshness Attack	50,000	50,000	100.0%
Total	100,000	100,000	100.0%

Table C.3: Adversarial Simulation Results (

N=100{,}000

Vectors)

Multi-Agent System Security

Multi-agent system security studies attacks, trust, privacy, and defenses in interacting AI agents, including language-model teams, robot swarms, and learned controllers.

See all 179 papers

MAGIQ: A Post-Quantum Multi-Agentic AI Governance System with Provable Security
May 7, 2026Sepideh Avizheh, Tushin Mallick, Alina Oprea +2AI Agent Governance Access Control
University of Calgary Calgary, Alberta, Canada · Northeastern University Boston, MA, USA
Securing Edge & Physical AI: Post-Quantum IoT Blockchain for Mission-Critical Cyber-Physical Agents
Oct 24, 2026Kürşat Çakal, Suat ÖzdemirCPS Security Multi-Agent System Security
Department of Computer Engineering, Hacettepe University, Ankara, 06800 Turkey
QSignAI: Quantum-Randomness-Seeded Identity Signatures at the Intersection of AI for Science and Science for AI
May 26, 2026Dongping Liu, Aoyu Zhang, Luyao Zhang
Amazon Web Services Hong Kong, China · Amazon Web Services Beijing, China · Duke Kunshan University Digital Innovation Research Center and Social Science Division Duke Avenue No.8, Kunshan, Suzhou, Jiangsu, China, 215316

May 7, 2026J/K move · Enter open · S save

MAGIQ: A Post-Quantum Multi-Agentic AI Governance System with Provable Security

Sepideh Avizheh, Tushin Mallick, Alina Oprea, Cristina Nita-Rotaru+1

University of Calgary Calgary, Alberta, Canada · Northeastern University Boston, MA, USA

AI Agent Governance Access Control Post-Quantum Cryptography Multi-Agent System Security

Open paper AskPDF arXiv

Our computing ecosystem is being transformed by two emerging paradigms: the increased deployment of agentic AI systems and advancements in quantum computing. With respect to agentic AI systems, one of the most critical problems is creating secure governing architectures that ensure agents follow their owners' communication and interaction policies and can be held accountable for the messages they exchange with other agents. With respect to quantum computing, existing systems must be retrofitted and new cryptographic mechanisms must be designed to ensure long-term security and quantum resistance. In fact, NIST recommends that standard public-key cryptographic algorithms, including RSA, Diffie-Hellman (DH), and elliptic-curve constructions (ECC), be deprecated starting in 2030 and disallowed after 2035. In this paper, we present MAGIQ, a framework for policy definition and enforcement in multi-agent AI systems using novel, highly efficient, quantum-resistant cryptographic protocols with proven security guarantees. MAGIQ (i) allows users to define rich communication and access-control policy budgets for agent-to-agent sessions and tasks, including global budgets for one-to-many agent sessions; (ii) enforces such policies using post-quantum cryptographic primitives; (iii) supports session-based enforcement of policies for agent-to-agent and one-to-many agent sessions; and (iv) provides accountability of agents to their users through message attribution. We formally model and prove the correctness and security of the system using the Universal Composability (UC) framework. We evaluate the computation and communication overhead of our framework and compare it with the state-of-the-art agentic AI framework SAGA. MAGIQ is a first step toward post-quantum-secure solutions for agentic AI systems.

Abstract

Multi-Agent System Security

Explore similar work

MAGIQ: A Post-Quantum Multi-Agentic AI Governance System with Provable Security

Securing Edge & Physical AI: Post-Quantum IoT Blockchain for Mission-Critical Cyber-Physical Agents

QSignAI: Quantum-Randomness-Seeded Identity Signatures at the Intersection of AI for Science and Science for AI