The insurance industry is tapping the brakes on artificial intelligence amid growing concerns about the scale of potential liabilities.
Three major US insurers, AIG, Great American, and WR Berkley, are now seeking regulatory approval to restrict their exposure to claims arising from AI agents and chatbots, marking one of the strongest indicators yet that the sector is bracing for an unprecedented level of risk.
Growing unease in the risk market
The rapid integration of AI systems into everyday business operations has left many insurers wondering: Who pays when an AI makes a mistake? The list of potential AI missteps continues to expand as more businesses integrate it into their operations, from customer-service chatbots giving poor financial advice to generative models creating harmful or defamatory content. Some executives worry that a single large-scale failure could generate claims running into the billions.
These concerns echo recent reporting from the Financial Times, which noted that insurers view AI-related failures as a growing source of potentially multibillion-dollar claims.
Unfortunately, this is no longer just theoretical. Insurers acknowledge that traditional liability frameworks were never designed for today’s autonomous, data-driven systems that interact with large numbers of people simultaneously. And as companies become increasingly reliant on AI-driven workflows, the risk of systemic failures affecting thousands of users simultaneously increases. For insurers, this introduces tail-risk scenarios far beyond those associated with conventional software issues.
AI litigation warning signs
Unease on the AI legal front has been evident in the surge of AI-related lawsuits and high-profile disputes involving generative models, with businesses deploying chatbots facing complaints about misinformation, privacy violations, and unintended financial consequences attributed to automated interactions. Companies are updating their AI guidelines, and regulators are increasingly scrutinizing AI outputs, pointing toward both a greater likelihood and potential cost of litigation.
Insurers reviewing these trends have determined that current policy language does not adequately protect them from catastrophic exposure in worst-case scenarios. Their push to amend liability terms reflects months of internal analysis and a belief that AI-driven claims could exceed existing reserves.
Changes that could reshape corporate protection
If regulators sign off on these limitations, businesses that rely on AI may soon find themselves with far less protection than expected. Some potential outcomes include companies facing higher premiums, narrower coverage, or the need to seek entirely new, specialized forms of AI liability insurance. While many smaller firms adopted AI to cut costs, these are the businesses that may be most affected as reduced coverage raises their operational risk.
Without clear guardrails around accountability and risk, the potential for AI innovation and adoption could slow, and smaller businesses, especially, may pull back from AI adoption as businesses become more cautious about potential costs. As risks rise, companies may have to adjust their operations to navigate the uncertain liabilities of the AI era.
Meanwhile, AI adoption in China is doing well. Alibaba’s new Qwen App has surpassed 10 million downloads within just a week of its public beta launch.