Insurers face unclear AI exclusions

Gallagher executives said on May 21 that insurers are still taking a wait-and-see approach to artificial-intelligence risk in directors and officers liability coverage, even as companies test AI tools more widely. Industry reporting indicates broad AI exclusions are not yet standard in D&O and professional-liability policies, leaving many disputes to turn on existing wording, definitions and facts of a claim. At the same time, cyber specialists are warning that AI is increasing the scale and sophistication of attacks faster than current market pricing may reflect. Legal commentators say that combination is pushing insurers, brokers and policyholders into a more detailed debate over how AI should be defined, priced and excluded. ### If insurers are worried about AI, why haven’t they broadly excluded it yet? Insurance Business reported on May 21 that Gallagher executives said insurers are closely watching AI-washing and other emerging liability risks, but have not broadly moved to exclude AI across D&O coverage. The publication said carriers remain cautious partly because many insureds are still experimenting with AI rather than deploying it at full scale. D&O insurers have handled new technology risks this way before: by relying first on existing policy language, disclosure standards and underwriting questions rather than rushing into market-wide exclusions. Insurance Business said the result today is that many policies are effectively silent on AI, which leaves room for future disputes over whether a claim fits within current grants of coverage or existing exclusions. (insurancebusinessmag.com) ### Where does the uncertainty show up first in a claim? Professional-liability and management-liability disputes can start with the wording already on the page. If a company is sued over alleged misstatements about an AI product, biased outputs, privacy violations or failures in oversight, the first fight may be over whether the loss is best characterized as a securities, professional-services, cyber or intellectual-property claim. National Law Review articles have said existing policies may respond to many AI-related losses, but the answer depends on the precise allegations and wording. (insurancebusinessmag.com) A National Law Review analysis published in July 2025 said few cyber policies expressly addressed AI, though a small but growing number of insurers were starting to add AI-specific definitions or terms. That means policyholders and insurers can face the same question from opposite sides: whether adding AI language clarifies coverage or narrows it. ### Why are cyber underwriters more vocal about pricing than exclusions? (natlawreview.com) Insurance Business reported on May 21 that cyber insurers said the market’s prolonged soft-pricing cycle is increasingly out of step with the threat environment as AI boosts the sophistication and scale of attacks. The publication said underwriters have not yet seen a single “Hurricane Katrina moment” for cyber, but some warned that soft rates could be masking accumulation risk. (natlawreview.com) Insurance Business had reported earlier, on May 7, that evolving threats and inconsistent wording were widening cyber coverage gaps as insurers tried to keep pace with AI-related exposures. In February, the same outlet said cyber insurers were rethinking limits, wording and underwriting models as AI-driven threats increased loss potential. (insurancebusinessmag.com) ### What kinds of AI disputes are insurers already bracing for? Insurance Business reported in October 2025 that AI had become a litigation flashpoint in D&O, with insurers focusing on corporate disclosures as non-technology companies came under pressure over AI claims. National Law Review said in a May 2026 article on Upstart that investor suits tied to AI model performance underscore the D&O implications when executives are accused of misrepresenting how an AI system works. (insurancebusinessmag.com) National Law Review also said in March 2024 that early AI litigation centered on privacy, copyright and defamation, while a later wave could reach a broader set of businesses adopting AI in products and services. That broader spread is one reason insurers and policyholders are now focusing more closely on whether AI should be handled inside existing lines or through more distinct wording. (insurancebusinessmag.com) ### What happens next for carriers, brokers and policyholders? The next step is likely to be more underwriting questions and more policy drafting rather than an immediate market-wide exclusion. Insurance Business reporting and National Law Review commentary both point to the same near-term pressure points: definitions of AI, the scope of professional services, cyber wording, disclosure risk and whether AI-related losses should sit inside existing coverage towers or be carved out more explicitly. (natlawreview.com) May 2026 reporting suggests the issue will now move through renewals, manuscript endorsements and future claims disputes. The clearest places to watch are D&O and cyber renewals, broker guidance on AI questionnaires, and court fights over how older policy language applies to newer AI allegations. (insurancebusinessmag.com)