Insurers Brace for AI Risks, Seek to Limit Coverage for Emerging Claims
Table of Contents
The insurance industry is rapidly moving to exclude artificial intelligence risks from corporate policies as businesses grapple with the potential for multibillion-dollar claims stemming from the rapidly evolving technology. Major players like AIG, Great American, and WR Berkley have recently petitioned US regulators for permission to offer policies that specifically exclude liabilities associated with companies deploying AI tools, including chatbots and automated agents.
The reluctance of insurers to provide comprehensive coverage comes as organizations across sectors rush to integrate cutting-edge AI, a trend already resulting in costly and embarrassing errors. These errors often involve AI models “hallucinating” – generating false or misleading information – and creating significant financial repercussions.
A Broad Exclusionary Approach
One proposed exclusion, submitted by WR Berkley, aims to bar claims arising from “any actual or alleged use” of AI, extending to any product or service incorporating the technology. In a filing responding to inquiries from the Illinois insurance regulator, AIG acknowledged that generative AI is a “wide-ranging technology” and the likelihood of future claims will “likely increase over time.” While AIG has filed for these exclusions, a company spokesperson told the Financial Times that they “have no plans to implement them at this time,” viewing approval as a means of retaining future flexibility. Both WR Berkley and Great American declined to provide comment.
The “Black Box” Problem
Insurers are increasingly wary of the unpredictable and opaque nature of AI outputs. “It’s too much of a black box,” explained a leading expert in cyber insurance at Mosaic. Even Mosaic, which offers coverage for some AI-enhanced software, has declined to underwrite risks associated with large language models like ChatGPT. “Nobody knows who’s liable if things go wrong,” stated Rajiv Dattani, co-founder of the Artificial Intelligence Underwriting Company, an AI insurance and auditing start-up.
Mounting Real-World Incidents
This shift in coverage comes amid a growing number of high-profile incidents linked to AI failures. Wolf River Electric, a solar company, is currently suing Google for defamation, seeking at least $110 million in damages after Google’s AI Overview feature falsely claimed the company was subject to legal action by Minnesota’s attorney-general.
Other cases include a tribunal ordering Air Canada to honor a discount falsely offered by its customer service chatbot, and a UK engineering firm, Arup, losing HK$200 million (US$25 million) after fraudsters used a digitally cloned executive to authorize fraudulent financial transfers during a video conference.
Systemic Risk and Aggregated Losses
The potential for widespread, correlated losses is a major concern for insurers. According to a senior official at Aon, the industry can absorb a $400 million to $500 million loss from a single company deploying flawed agentic AI that delivers incorrect pricing or medical diagnoses. However, “What they can’t afford is if an AI provider makes a mistake that ends up as a 1,000 or 10,000 losses — a systemic, correlated, aggregated risk.”
AI hallucinations typically fall outside the scope of standard cyber insurance, which focuses on security and privacy breaches. While “errors and omissions” policies may offer some coverage, insurers are introducing carve-outs to limit their exposure.
Insurers are attempting to clarify legal ambiguities through policy “endorsements” – amendments to existing policies. However, brokers caution that these endorsements often result in reduced coverage. For example, an endorsement by QBE, mirroring actions by other insurers, extended coverage for fines under the EU’s AI Act, but capped payouts at just 2.5% of the total policy limit. QBE stated they were “addressing the potential gap [in AI-related risk] that may not be covered by other insurance policies.”
Zurich-based Chubb has agreed to cover some AI risks in negotiations, but has explicitly excluded “widespread” incidents impacting multiple clients. Chubb declined to comment. Other insurers are introducing add-ons for narrowly defined AI risks, such as chatbot malfunctions.
The Inevitable Legal Battles
Insurance brokers and legal experts anticipate a surge in court battles as AI-driven losses increase. “It will probably take a big systemic event for insurers to say, hang on, we never meant to cover this type of event,” predicted Aaron Le Marquer, head of insurance disputes at law firm Stewarts. The evolving landscape of AI risk suggests a period of significant uncertainty and legal challenges lies ahead for both insurers and businesses alike.
