The rapid adoption of artificial intelligence across industries is presenting an unprecedented challenge for the insurance sector, with major players now deeming AI too risky to cover. Leading insurers, including AIG, Great American, and WR Berkley, are actively seeking permission from U.S. regulators to exclude AI-related liabilities from their corporate policies. This move, as reported by the Financial Times, signals a growing apprehension within an industry whose core business is managing risk.

The "Black Box" Problem

A primary concern for underwriters is the opaque nature of AI models. One underwriter described AI outputs to the FT as "too much of a black box," highlighting the difficulty in understanding how these complex systems arrive at their conclusions. This lack of transparency makes it incredibly challenging to assess, price, and ultimately insure against potential failures or unintended consequences.

A Growing List of AI Mishaps

The insurance industry's caution is well-founded, given a series of high-profile incidents involving AI:

  • In March, Google's AI Overview feature falsely accused a solar company of legal troubles, leading to a $110 million lawsuit.
  • Last year, Air Canada was forced to honor a discount that its chatbot erroneously invented for a customer.
  • Fraudsters successfully used a digitally cloned voice of a senior executive to steal $25 million from London-based design engineering firm Arup during a deepfake video call that appeared entirely legitimate.

These incidents underscore the real-world financial and reputational damages that AI errors can inflict.

The Threat of Systemic Risk

While individual large payouts are a concern, what truly "terrifies" insurers is the prospect of systemic risk. This refers to the potential for thousands of simultaneous claims triggered by a single, widely used AI model making a critical error. As an executive from Aon, a global professional services firm, articulated, insurers are equipped to handle a $400 million loss to one company. However, they are ill-prepared for an "agentic AI mishap" that could simultaneously generate 10,000 losses across multiple clients.

The move by major insurers to seek regulatory exclusion for AI liabilities highlights a critical juncture in the integration of artificial intelligence into daily operations. As AI technology continues to evolve, the challenge of managing its inherent risks will increasingly fall under intense scrutiny from both regulators and the broader business community.