Insurers are wrestling with a hidden problem called silent AI - artificial intelligence embedded into existing products, services and processes that creates exposures policyholders and underwriters may not even know exist. The debate centers on whether to address these risks within current coverage or to explicitly exclude them, even as claims for AI-related losses start flowing in against standard commercial policies.
How silent AI shifts liability across different users
Hayley Budd, head of innovation at U.K.-based underwriting firm Atrium, said the type of AI involved directly affects the liability profile. Generative AI can produce incorrect advice or flawed outputs, while agentic AI takes actions that cause damage. "Whether that is booking a table for dinner or a holiday or authorizing a payment, purchasing goods, making employment decisions, or even getting to a point of handling claims and paying out claims without a human being involved, the liability is switching from an output to what the AI actually does," Budd said during a recent webcast.
The exposure also depends heavily on who holds the policy. A base model provider faces a fundamentally different risk than an e-commerce company deploying off-the-shelf chatbots. Budd explained that even with identical underlying technology, "the liability profile is going to change depending on what you're doing with the AI. AI liability isn't like one definition of one risk. It's a collection of liabilities that arise from different points, different uses, different people."
For insurance professionals navigating this shifting terrain, structured education on AI for Insurance can clarify how different AI applications map to existing coverage frameworks.
The exclusion trend in commercial policies
Not everyone sees a need for new affirmative coverage that names AI risks explicitly. James Cooper, partner at Norton Rose Fulbright LLP, argued the fundamentals remain unchanged. "The liability is the same as it always has been," he said. "Is it negligence? Whose fault is it? Can we pass liability on to somebody else?"
Yet commercial policies for financial institutions and directors and officers coverage are already moving in the opposite direction - adding language that excludes AI risks. Cooper pointed out a practical problem with either approach: identifying whether AI was truly the cause behind a loss. "When machines can do things without a human's interaction, how will first-party policies be able to respond to that?" he said.
Limited capacity meets growing demand
Despite the definitional challenges, demand for dedicated AI coverage persists. Nick Gibbs, class leader at Apollo, a Skyward Group insurance platform operating through Lloyd's of London, described a market that is watching and waiting. "It remains in the background, against a background environment of people cautiously monitoring it," Gibbs said. "There needs to be more activity in the market for there to be greater appetite around it."
A handful of Lloyd's syndicates and U.S. managing general agents now offer dedicated AI liability products, but Budd cautioned that these remain narrow in scope. Limits are often modest, with hard aggregate caps that may not match a client's actual exposure. For companies using AI in limited ways, those limits might suffice today. That changes quickly as AI becomes embedded in critical business processes or large-scale agent systems. "The market is going to need significantly more capacity than what is currently available, both from a primary insurance and a reinsurance standpoint as well," Budd said.
Why this matters for insurance professionals
Silent AI is not a future problem - it is already creating coverage disputes under existing policies. Underwriters, claims adjusters, and brokers need to examine their books for AI exposures embedded in standard commercial lines, particularly as agentic systems take actions without human oversight. The gap between available AI-specific capacity and rising demand also signals a product development opportunity for carriers willing to define the risk clearly and build appropriate limits.
Your membership also unlocks:
