Insurance companies seek AI exclusions

Insurance companies in the United States are requesting approval to include exclusions related to artificial intelligence in their policies. Over 80 percent of U.S. state insurance commissioners have already approved these requests in states like Florida, Connecticut, and Maryland. Major companies such as AIG, Great American, and WR Berkley are working with regulators to implement these changes.

Startups develop AI-specific insurance

A small group of startups and specialized insurers are creating insurance products designed specifically for AI-related risks. For example, Klaimee, a startup backed by Y Combinator, raised $5.5 million in July to audit autonomous AI agents and offer liability coverage. HSB, a subsidiary of Munich Re based in Connecticut, has been selling AI liability insurance for small and medium-sized businesses since March.

AI insurance covers multiple risks

The term "AI insurance" can include various types of coverage, such as losses suffered by the company, compensation for third-party claims, and compensation for AI models that fail to meet performance standards. These conditions are not interchangeable, and companies are encouraged to examine the exact wording of their policies and assess them based on specific claims.

Underwriting challenges with AI

Risk management experts note that there is not enough historical data to fully assess AI-related business risks. Traditional insurance underwriting relies on past data to predict future losses, but AI evolves too quickly for these models to keep up. Data from past losses may not reliably predict future exposure, making it difficult to assess risk accurately.

Lack of actuarial data on AI claims

There is no established actuarial history of claims for AI systems, and the current underwriting approach is similar to the early days of cyber insurance. A recent IBM study found that only 24 percent of enterprise AI projects include a security component, despite 82 percent of executives believing security is essential for AI implementation.

Growth in AI-related disputes

AI-specialized insurance startup Testudo reports a 137 percent increase in AI-related disputes compared to last year. However, the few insurers offering dedicated AI coverage reveal a gap in the market. Mark Walters, CEO of Alliance Risk, says he has not seen enough public claims related to AI to confirm the effectiveness of these policies.

Legal and contractual pressures

Executives are increasingly anxious about AI risks, and new legal clauses in B2B contracts are driving demand for insurance. Security and legal teams now include specific indemnification and insurance requirements related to AI in supplier contracts. Compliance requirements are also pushing companies to act, with the ISO approving new exclusion forms related to generative AI.

Regulatory focus on liability

Several insurers are adding these exclusion forms during policy renewals without adequately explaining their impact on liability. Once implicit coverage is removed, buyers seek explicit coverage. In the United Kingdom and the European Union, the GDPR and the AI Act protect consumer privacy but do not cover companies' liability for AI-related harms.