Midsize U.S. insurers will factor AI risks into 60-80% of new policies and renewals across errors and omissions, directors and officers, employment practices liability, and cyber insurance lines by 2028, according to proprietary research released by ScienceSoft on September 10, 2026. The forecast signals a shift in underwriting practice that will touch nearly every commercial policy these carriers write, even as dedicated AI insurance products remain a sliver of the market.
What's driving the change
The research points to a 262% rise in publicly documented AI incidents between 2022 and 2025. Businesses are showing strong interest in protection against liabilities tied to algorithmic decisions, automated hiring tools, and AI-generated content. Insurers are responding by making AI treatment more explicit through affirmative policy wording, AI exclusions, specialized endorsements, and a small but growing number of dedicated AI insurance products.
Despite that momentum, ScienceSoft expects most midsize carriers to continue covering AI risks through existing lines of business rather than standalone policies. The firm projects AI-specific insurance will grow from $40 million in 2024 to $4.8 billion by 2032 - an approximately 80% compound annual growth rate - yet still account for only around 0.34% of commercial property and casualty premiums by that year.
Underwriting adapts faster than coverage
The research finds insurers are beginning to assess not only whether businesses use AI, but how those systems are governed, what level of autonomy they have, and what controls organizations have in place. These factors will increasingly influence premiums, coverage conditions, and risk control requirements through 2028.
Unclear liability attribution, accumulation risk, limited loss history, and regulatory uncertainty continue to slow development of dedicated AI for Insurance coverage. The liability questions are especially thorny: when an AI system causes harm, responsibility may fall on the developer, the deployer, or both. Traditional insurance frameworks were not built for that ambiguity.
Who feels the impact
Beyond insurers themselves, the research examines consequences for brokers, commercial insurance customers, AI software providers, and regulators. ScienceSoft suggests insureds may increasingly require broker services as AI-specific insurance develops, and that AI vendors may become one of the main customer groups for AI liability insurance. Cyber underwriters, already navigating a fast-changing threat environment, will need to account for AI-enabled attacks and AI system vulnerabilities - a concern that overlaps with priorities for AI for Cybersecurity Analysts.
Why this matters for insurance professionals
The 60-80% figure is not a distant forecast. It means that within two years, underwriters who have not built frameworks for evaluating AI governance, autonomy levels, and control environments will be making decisions on incomplete information for a majority of their books. Brokers who cannot explain how standard E&O or cyber policies treat AI-related claims will lose credibility with commercial clients who are already asking those questions. The market is not waiting for a perfect standalone AI product - it is adapting existing tools now.
Your membership also unlocks:
