This report examines how artificial intelligence (AI) is being adopted across supply chains, including logistics, transport, warehousing, and inventory management. It explores how AI may create new operational dependencies, resilience challenges, and insurance risks, including correlated losses across firms, and considers implications for operators, insurers, and regulators.

Operational reliance on AI in supply chains and emerging insurance risks

Elie Alhajjar, Youmna Hashem, Alec Ross, Sasha Romanosky

ResearchPublished Sep 2, 2026

Artificial intelligence (AI) is becoming embedded in the digital infrastructure through which goods are forecast, routed, documented, stored, monitored, and delivered. In transportation, logistics, warehousing, and inventory management, AI is being used for forecasting, route optimisation, predictive maintenance, document processing, warehouse robotics, and anomaly detection. These applications may improve efficiency, visibility, and responsiveness, but they also create new forms of operational reliance on data, models, software vendors, cloud services, and automated decision rules.

This report examines how that operational reliance may translate into insurance risk, particularly where AI-related failures affect multiple firms at the same time. The focus is on transportation and logistics, and on warehousing and inventory management, with particular attention to aggregation risk: the possibility that shared AI weaknesses, common service dependencies, cyberattacks, or regulatory shocks generate correlated losses across policyholders and insurance lines.

The analysis combines a review of academic and industry literature, a desk-based review of AI-related incidents and lawsuits, stakeholder interviews across operational, legal, insurance, and technology backgrounds, and structured scenarios designed to test how AI-related losses could spread across firms and portfolios. The findings suggest that AI adoption in supply chains is likely to expand over the next five years, but unevenly. Lower risk uses such as monitoring, forecasting, and recommendations are likely to spread faster than automated execution. At the same time, insurers, operators, and regulators still lack settled approaches to identifying, governing, and underwriting AI-related supply chain risk.

Key Takeaways

AI adoption in supply chains is likely to grow, but unevenly

The strongest near-term growth is expected in lower-risk applications that can be added to existing digital workflows, such as monitoring, forecasting, optimisation, predictive maintenance, and document processing. More autonomous or execution-oriented uses are likely to expand more slowly and selectively because they raise greater safety, liability, integration, and business continuity concerns.

AI-related failures can create hidden dependencies and correlated insurance losses

Firms that appear independent at the underwriting level may rely on the same model provider, cloud platform, warehouse robotics system, mapping service, logistics software, or data feed. This creates aggregation risk: a shared AI weakness, service outage, cyberattack, or regulatory shock could generate related losses across multiple policyholders and insurance lines at the same time.

Governance and insurance responses are still evolving

Current AI use in supply chains remains more assistive than autonomous in many settings, with continued reliance on human judgement in higher-stakes contexts. At the same time, responsibility for AI-related failures remains difficult to allocate across operators, developers, vendors, and integrators. Insurers are beginning to adapt through changes to existing policy wordings, exclusions, and case-by-case assessments, but no settled approach to underwriting or covering AI-related supply chain risk has yet emerged.

Recommendations

  • Supply chain operators and AI and technology providers should establish baseline controls before AI systems become operationally indispensable, including AI-use inventories, meaningful human oversight, provenance tracking, runtime monitoring, staged deployment, and tested fallback procedures.
  • Insurers and reinsurers should incorporate AI into underwriting and accumulation management rather than treating AI use as a binary exposure. This should include asking where AI is used, what operational authority it has, which vendors and models are involved, whether updates are tested, and what human review, monitoring, and fallback arrangements exist.
  • National regulators should distinguish manageable portfolio aggregation from genuinely catastrophic AI risk and assess whether existing insurance and regulatory tools are adequate. Where private markets may not be able to absorb extreme correlated AI-related losses, regulators should examine options such as a public-private insurance mechanism, government backstop, or other coordinated response.

Topics

Document Details

Citation

Chicago Manual of Style

Alhajjar, Elie, Youmna Hashem, Alec Ross, and Sasha Romanosky, Operational reliance on AI in supply chains and emerging insurance risks. Santa Monica, CA: RAND Corporation, 2026. https://www.rand.org/pubs/research_reports/RRA5093-1.html.
BibTeX RIS

Research conducted by

This research was sponsored by the AI Security Institute and conducted by the Science and Emerging Technology Program of RAND Europe.

This publication is part of the RAND research report series. Research reports present research findings and objective analysis that address the challenges facing the public and private sectors. All RAND research reports undergo rigorous peer review to ensure high standards for research quality and objectivity.

This document and trademark(s) contained herein are protected by law. This representation of RAND intellectual property is provided for noncommercial use only. Unauthorized posting of this publication online is prohibited; linking directly to this product page is encouraged. Permission is required from RAND to reproduce, or reuse in another form, any of its research documents for commercial purposes. For information on reprint and reuse permissions, please visit www.rand.org/pubs/permissions.

RAND is a nonprofit institution that helps improve policy and decisionmaking through research and analysis. RAND's publications do not necessarily reflect the opinions of its research clients and sponsors.