As artificial intelligence becomes increasingly embedded in business operations, major insurers are sounding the alarm over its growing risks—so much so that they’re seeking permission from U.S. regulators to exclude AI-related liabilities from standard corporate insurance policies. Companies including AIG, Great American, and WR Berkley are urging regulators to allow them to carve out coverage for AI-driven harms, citing the technology’s unpredictable nature and lack of transparency. Insurers are particularly unsettled by what they describe as the “black box” nature of many AI models—systems whose decision-making processes are opaque even to their creators. This uncertainty makes it nearly impossible to assess risk accurately or assign liability when things go wrong. The concerns are grounded in real-world incidents. In March, Google’s AI Overview falsely claimed a solar energy company had faced legal issues, sparking a $110 million lawsuit. Earlier, Air Canada was forced to honor a discount offered by a chatbot that had no authority to issue such promotions. Perhaps most alarming, fraudsters used a deepfake audio and video simulation of a senior executive to trick employees at Arup, a London-based engineering firm, into transferring $25 million in a single video call that appeared entirely legitimate. What truly worries insurers isn’t the potential for a single massive payout—it’s the possibility of cascading, simultaneous claims. A flaw in a widely used AI system could trigger thousands of losses across different industries at once. As one Aon executive noted, insurers can manage a $400 million loss from a single company. But they cannot survive a scenario where an agentic AI—capable of autonomous decision-making—causes widespread harm across thousands of businesses in a short time. That kind of systemic risk threatens the entire insurance market. With AI adoption accelerating and regulatory frameworks lagging behind, the industry is pushing back, demanding clarity and protection. Without clear rules and accountability, insurers say they may be forced to pull back from covering AI-related risks altogether—potentially leaving companies exposed and slowing the responsible development of the technology.