Insurance Market Pulls Back from AI Risks as Corporate Liability Questions Mount

Insurers Wary of Corporate Liability Amid Rising AI Risks

The insurance market is retreating from AI-related liabilities as companies accelerate artificial intelligence adoption without clear coverage frameworks. W. R. Berkley, a major insurer, has introduced exclusions in its D&O (directors and officers), E&O (errors and omissions), and Fiduciary Liability products to exclude coverage for "any actual or alleged use, deployment, or development of Artificial Intelligence."1

1

During the company's Q4 2025 earnings call, CEO W. Robert Berkley emphasized underwriters need to "fully understand that risk so we can control it, select it, and price for it."1

1

RAND Corporation Documents Scope of AI-Related Liabilities

A newly issued RAND Corporation report reveals that "AI-related harms are already emerging, including incorrect or misleading outputs, deepfakes, privacy violations, intellectual property disputes, fraud, product defects, and discriminatory decisions."1

1

The Artificial Intelligence Incident Database (AIIDB) currently lists 713 incidents drawn from more than 6,000 reports. The categorical breakdown includes misinformation and manipulation (586 incidents), deepfakes and synthetic media (346), hallucination and factual error (215), harmful content (92), privacy and data leak (58), bias and discrimination (47), and copyright and IP (20).1

1

Additionally, approximately 250 US lawsuits related to AI are currently active, largely concerning copyright and IP but also touching on privacy, fraud, negligence, and discrimination.1

1

Enterprise Insurance Market Struggles to Underwrite AI Risks

The enterprise insurance market faces a fundamental challenge with AI insurance: traditional risk diversification doesn't work. Unlike houses that burn down independently, thousands of businesses may rely on the same foundation models, cloud infrastructure, or agent framework simultaneously. A defect or vulnerability upstream could generate losses across many supposedly independent policyholders at once.2

2

A malfunctioning agent could create professional liability in one situation, cyber losses in another, and potentially directors-and-officers issues if management's deployment becomes part of litigation.2

2

In January 2026, Verisk/ISO, whose standardized forms appear in more than 80 percent of US property and casualty policies, introduced optional language carriers can adopt to exclude bodily injury, property damage, and other harms arising from generative AI.1

1

Agentic AI's Hidden Ceiling Created by Insurance Coverage Gaps

The deployment of agentic AI faces practical limits determined not by technical capability but by what insurers will cover. Nearly 7% of enterprise CFOs in the United States had deployed agentic AI in live finance workflows as of September 2025, while an additional 5% were running pilots.2

2

However, insurance coverage for AI remains fragmented. Companies seeking coverage may increasingly need to demonstrate how models are tested, what systems agents can access, what transactions they can execute, whether humans approve consequential actions, how AI incidents are monitored, and whether an agent's actions can be reconstructed afterward.2

2

Insurers Emerging as Private Regulators Through Coverage Requirements

The insurance market is positioning itself as a de facto regulator of AI deployment through coverage requirements. Boards, lenders, procurement departments, and counterparties routinely require organizations to demonstrate adequate coverage before assuming particular risks. An enterprise may technically be able to let an agent autonomously move $10 million, modify production code, or negotiate contracts, but its insurer may simply decline to cover that configuration without additional controls.2

2

Alaap Shah, member of the firm at Epstein Becker Green, noted that "there are existing bodies of law that, while not passed or promulgated for the reason of AI, are still applicable to AI solutions."2

2

Path Forward Requires Standardized Taxonomy and Governance Frameworks

RAND argues that policy researchers, brokers, carriers, and reinsurers need to develop a common standardized taxonomy to track AI incidents and claims. The think tank wants state regulators to push for an AI Coverage Notice so everyone is clear on what's covered and what isn't.1

1

Not all insurers are deciding against covering AI-related liabilities. Coverage gaps are being filled by new and existing companies that believe they have a handle on the risk calculations.1

1

The practical ceiling on autonomy will be set not by what artificial intelligence can do, but by what risk markets will finance, creating new governance frameworks that balance innovation with accountability.