AI Lawsuits Reveal Insurance Gaps as Autonomous Systems Outpace Conventional Coverage

AI lawsuits expose gaps in conventional insurance, says report

AI is exposing a blind spot in the insurance industry, with many insurers on the hook for risks embedded in policies that were never designed to cover increasingly autonomous AI systems. A report by the Artificial Intelligence Underwriting Company, co-authored with researchers from Anthropic and OpenAI, says that more than 90 per cent of insurers' exposure to AI agents sits in "silent" cover embedded in conventional policies, leaving risks largely unpriced and often invisible to insurers themselves. The review says rising litigation reflects a broader shift as risks concerning AI systems move beyond chatbots producing erroneous content to "agents" that take actions. That development may expose businesses to claims ranging from professional negligence, cyber attacks and wrongful death. Some industry executives, such as brokers who represent companies buying insurance, have said such warnings are overdone and form an effort to sell a range of new AI insurance products. But Rajiv Dattani, co-founder of AIUC, said legal liabilities had become a constraint on wider use of AI by large companies. "Businesses cannot adopt AI unless they know the risk has been quantified and managed," he said. A series of recent cases has brought the legal issues into focus. Google is defending a lawsuit seeking at least $110mn after its AI Overviews system allegedly defamed US solar company Wolf River Electric, while Air Canada was ordered to honour a discount invented by its customer service chatbot. Last year, UK engineering group Arup lost HK$200mn (US$25mn) after fraudsters used AI-generated deepfakes of senior executives to persuade an employee to transfer funds. The report says such cases foreshadow much larger liabilities as AI systems become more autonomous. It estimates that while a major AI disaster might cause about $100bn in direct damage, the wider economic impact could run into the trillions of dollars if insurers withdraw cover, businesses slow AI adoption and investors retreat from the sector. The authors compare the risk to the collapse of the terrorism insurance market after the September 11 attacks, when more than $40bn of insured losses prompted insurers to retreat from the market, stalling construction projects and disrupting commercial aviation until governments stepped in to provide backstops. Kevin Kalinich, head of intangible assets at broker Aon and a co-author of the report, said AI could ultimately produce even larger "aggregated, systemic, correlated" losses if a malfunctioning model or co-ordinated attack simultaneously affected hundreds or thousands of companies. Such an event could also expose big major gaps in existing policies, prompting disputes over whether insurers were ever liable for AI-related losses, while AI is already amplifying existing risks such as deepfake-enabled fraud that often falls outside conventional cyber or liability cover, Kalinich added. Rather than simply excluding AI risks from existing policies, the report says insurers should develop dedicated cover for increasingly autonomous AI systems, supported by new underwriting standards and technical auditing that keeps pace with models whose capabilities evolve far faster than traditional actuarial data. "As these systems get more powerful, the incidents are getting more severe," Dattani added. "In 2022, the worst these systems could do was hallucinate a fake refund policy. Today, it's wrongful death and unauthorised practice of law." Cyber risk is a particular pressure point for insurers, where advances in technology have outpaced traditional coverage. The release of Anthropic's Claude Mythos model demonstrated the potential for new models to pose sudden, widespread risks that ripple across industries. "The discovery that Mythos displays a step-change in cyber capabilities sets a pattern we can expect to recur: powerful new capabilities emerging unpredictably, each one reshaping the risk landscape in ways that are difficult to foresee," said Matthew Botvinick, who leads work at Anthropic on AI and the rule of law, and co-authored the report. Tim Zawacki, insurance analyst at S&P, said insurers had been slow to strip out the risk of silent AI coverage from policies, for fear of losing business. But he added: "It will take a catalyst like a loss or a high-profile incident for insurers to exclude these AI risks en masse."

Insurance Industry May Be Unprepared for Agentic AI Risks | PYMNTS.com

Findings from the Artificial Intelligence Underwriting Company (AIUC), the subject of a Tuesday (July 14) Financial Times (FT) report, show that more than 90% of insurers' exposure to artificial intelligence (AI) is in "silent" cover that is part of conventional policies. That means risks are "largely unpriced" and in many cases unnoticed by insurers, said the report, co-authored with researchers from Anthropic and OpenAI. The researchers also say rising litigation indicates a larger shift, with AI risks going from chatbots generating bogus information to AI agents taking action. This could leave businesses facing claims including professional negligence and wrongful death. While some industry figures say these warnings are overblown, and part of a campaign to push new AI insurance products, AIUC Co-Founder Rajiv Dattani said legal liabilities had held back broader AI use at larger companies. "Businesses cannot adopt AI unless they know the risk has been quantified and managed," he said. The FT cites some recent legal cases that highlight the issue. For example, Google is facing a suit seeking at least $110 million after its AI Overviews system allegedly defamed solar company Wolf River Electric, while a judge ordered Air Canada to honor a discount invented by its AI customer service chatbot. The AIUC report argues these cases portend even bigger liabilities as AI systems grow more autonomous. It estimates that although an AI catastrophe could cause around $100 billion in direct damages, the larger economic fallout could be in the trillions of dollars if insurers pull coverage, businesses cool on AI adoption and investors shift away from the industry. As covered here earlier this month, the insurance industry and regulators have begun pricing AI hallucination risks. For example, Lloyd's of London last year introduced an insurance product specifically covering losses related to artificial intelligence hallucinations. FINRA's 2026 Annual Regulatory Oversight Report noted hallucinations as a specific compliance risk for broker-dealers, cautioning businesses to establish procedures for AI agents that may act beyond the user's intended scope. Cases like the Air Canada incident, that report added, are not about a failure of AI to understand the questions asked of it. "It is a failure of the system to recognize that it does not know the answer, and to say so," PYMNTS wrote. "When the invented answer concerns a company's pricing, policy or refund terms, courts, regulators and insurers now agree: The company owns what its model said."