Insurers Win Approval to Exclude AI Liability From General Policies — And a New Market Is Filling the Gap

By Hector Herrera | May 11, 2026 | Finance

State regulators have approved more than 80% of requests from major insurers to exclude AI-related damages from standard general liability policies, with exclusions already in force in Florida, Connecticut, and Maryland. The result is a coverage gap that a new category of standalone AI liability insurance is racing to fill — and a financial signal that the insurance industry now formally treats AI risk as a distinct and serious exposure class.

When insurers stop covering something under standard policies, it is not a bureaucratic footnote. It is the insurance industry's most definitive statement about risk.

What Changed and Where

According to Roots.ai's May 2026 industry analysis, Berkshire Hathaway, Chubb, and Travelers are among the carriers that have received regulatory approval to carve AI-related damages out of general liability policies. The approvals are not uniform across all states, but Florida, Connecticut, and Maryland have active exclusion language already in effect.

The exclusions cover damages arising from AI system decisions, AI-generated content causing harm, AI-driven discrimination, and AI errors in automated workflows. Under the new language, a company whose AI hiring tool discriminates against protected classes, or whose AI-generated medical content harms a patient, would find that general liability does not respond — even if the policy would have covered similar harm caused by a human employee.

The Standalone AI Market Is Emerging Fast

The gap created by exclusions is not going unfilled. Munich Re is among the established reinsurers developing dedicated AI liability products. A cluster of specialized startups — Corgi, Armilla, and Embroker — are competing to build the infrastructure for AI risk underwriting from scratch.

Current standalone AI liability products offer limits in the $2 million to $50 million range, with pricing that varies significantly based on:

• Deployment context. AI in healthcare or financial services commands substantially higher premiums than AI in marketing automation.
• Auditability. Companies with documented model governance — bias testing, accuracy benchmarking, audit trails — are receiving better pricing than those without.
• Third-party AI reliance. Companies using AI-as-a-service through vendors face different underwriting logic than those running proprietary models.

The market is nascent and pricing is inconsistent. Brokers who specialize in tech errors and omissions (E&O) insurance are currently the primary distribution channel, but dedicated AI risk practices are forming at major brokerage firms.

The Deepfake Claims Problem

Parallel to the liability exclusion trend, Roots.ai documents a rising volume of AI-generated deepfake imagery appearing in property insurance claims and medical insurance submissions. Fraudsters are using AI image generation to fabricate property damage, manipulate repair evidence, and generate synthetic medical documentation.

This creates a direct cost pressure on carriers — one that makes AI risk more visible to actuarial teams and accelerates the push to separate AI exposure from general underwriting assumptions. The fraud vector is still relatively small in absolute claim volume, but it is growing faster than most carriers' fraud detection systems were designed to handle.

What This Means for Businesses

If you are buying liability insurance and using AI in your operations, the exclusion landscape requires immediate attention. Many companies are unaware that AI-related claims may not trigger coverage under policies they have held for years. The exclusion language is being added at renewal — often in fine print that does not get flagged by risk managers focused on headline coverage limits.

For companies using AI in high-stakes workflows — healthcare, financial advice, legal services, hiring decisions — the exposure is most acute. A healthcare company whose AI diagnostic tool provides an incorrect recommendation now faces liability risk that a standard policy may explicitly exclude.

Key questions to ask your broker now:

• Does our current general liability policy contain AI exclusion language? When was it added?
• Is standalone AI liability coverage available for our specific deployment context?
• What documentation requirements would position us for better pricing on an AI-specific policy?

For risk managers, the emergence of standalone AI policies is a useful forcing function. Applying for AI liability coverage requires an organization to document its AI governance practices in ways that many companies have not yet formalized. The underwriting process is, in effect, a risk audit.

The Broader Signal

The insurance industry's formal separation of AI risk from general risk categories is significant beyond the premium market. Insurance pricing and exclusion decisions are among the most data-driven risk assessments that exist in the private sector. When major carriers determine that AI exposure does not fit within the actuarial assumptions underlying general liability, it signals that AI-related damage events are frequent enough, severe enough, or unpredictable enough to require dedicated modeling.

That determination does not emerge from a press release or a policy position paper. It emerges from claims data. The exclusions are the industry's evidence-based conclusion.

What to Watch

• State-level approval rate for exclusions. If approvals continue above 80%, exclusion will become the industry standard across most major commercial markets within 12–18 months.
• Standalone policy pricing convergence. The $2M–$50M range reflects an immature market without reliable loss history. As claims data accumulates, expect pricing to become more precise — and potentially more expensive for high-risk categories.
• Regulatory response. Some states may mandate that insurers offer AI liability coverage rather than permitting exclusion-only approaches. Watch for model legislation from the National Association of Insurance Commissioners (NAIC), which has active AI working groups.
• Deepfake fraud detection investment. Carriers absorbing growing deepfake-enabled claim fraud are investing in AI detection tools — creating the somewhat ironic situation of AI being used to detect AI fraud.

The coverage gap is real and growing. The market filling it is nascent but moving quickly.

Source: Roots.ai — May 2026 Insurance AI Trends

NexChron articles on financial topics are for informational purposes only and do not constitute financial, legal, or insurance advice. Consult a licensed professional for guidance specific to your situation.