Major insurers are moving to draw a clearer line around AI risks in corporate insurance policies. According to the Financial Times, several large insurers, including AIG, Great American, and WR Berkley, have filed requests with U.S. regulators to exclude AI-related risks from their corporate insurance policies.
The filings show how quickly generative AI has become a concern for the insurance industry. Chatbots, AI agents and AI-generated answers can create new kinds of claims, and insurers are warning that the potential exposure could reach billions of dollars.
What the insurers are asking for
The central request is straightforward: insurers want permission to keep certain AI-related risks outside corporate insurance policies. The source article says several large insurers have made filings with U.S. regulators, with AIG, Great American and WR Berkley named among them.
WR Berkley reportedly proposed an exclusion that would apply to any claim resulting from the use of AI in any form. That wording matters because it would not be limited to one category of AI tool, one type of business use or one narrow failure mode. It would focus on whether the claim resulted from AI use.
AIG told Illinois insurance regulators that generative AI is a broad and far-reaching technology. It also said related claims are likely to increase in the future. That position points to the core issue for insurers: they are not only looking at what has already happened, but at the possibility that similar claims could become more common as companies use these tools more widely.
Why generative AI changes the risk calculation
Generative AI systems like chatbots and AI agents are different from many older business tools because they can produce answers, summaries or actions that look authoritative while still being wrong. The source article does not present this as a theoretical issue alone. It points to recent lawsuits and disputes as evidence that AI-generated outputs can create real liability concerns.
For insurers, the problem is not just whether one company faces one claim. The concern is that a single type of AI failure could affect many companies at once. If the same AI provider, system or pattern of use produces a mistake across many customers, claims may not arrive as isolated events.
That is why the phrase correlated claims is important. Kevin Kalinich, a managing director at Aon, said the insurance industry could handle a single $400 million loss but not 1,000 or 10,000 correlated claims caused by an error from one AI provider. In plain terms, the danger is concentration: many claims tied to the same source of failure.
The lawsuits insurers are watching
The source article highlights two examples that help explain why insurers are paying attention.
- Wolf River Electric sued Google for at least $110 million, alleging that the company’s AI-generated overview spread false statements.
- In another case, a court ordered Air Canada to honor a discount price its customer service chatbot had invented.
These examples involve different settings, but they point to the same underlying concern. AI-generated content can affect what customers, businesses or the public believe. When that content is wrong, the result can become a legal and financial problem.
The Google case centers on an AI-generated overview and alleged false statements. The Air Canada case centers on a customer service chatbot and a discount price that the chatbot invented. Both examples show why insurers are looking closely at chatbots, AI agents and generated answers as possible sources of liability claims.
What this means for companies
The filings do not mean every corporate insurance policy has already changed. The source article says the insurers have filed requests with U.S. regulators. That distinction matters: the requests show what insurers are seeking, not a completed industry-wide outcome.
Still, the direction is clear. Large insurers are trying to limit how much AI-related liability they may have to absorb through corporate policies. If exclusions are allowed, companies using generative AI may need to pay closer attention to which AI risks are covered and which are not.
The issue is especially important because the proposed exclusions are tied to the use of AI itself. A business may use AI in customer service, summaries, internal workflows or other tasks, but the source article’s key point is broader: insurers are concerned that AI-related claims could grow and become difficult to price or contain.
For now, the insurance industry is signaling caution. Generative AI may help companies automate or expand work, but insurers are warning that it can also produce claims that are large, repeated and connected. The filings from AIG, Great American and WR Berkley show that AI risk is becoming a policy issue, not just a technology issue.
The bigger insurance question
The dispute comes down to who carries the financial consequences when generative AI systems cause harm. If insurers exclude AI-related risks from corporate insurance policies, more of that exposure may sit outside traditional coverage. If they do not, insurers worry about billions of dollars in liability claims and the possibility of many correlated losses.
That tension is likely to shape how companies, insurers and regulators discuss AI risk. The source article shows that the debate has already moved into regulatory filings, and that major insurers are pushing to define the boundaries before claims become more frequent.