2026-08-12 · Source: Risk & Insurance Magazine
Summary in 3 Points • Resilience's report found no AI-specific attack losses in H1 2026 • Human error accounted for 85.3% of cyber losses in H1 2026 • Claims frequency rose to 45.9 per 100 policies in H1 2026 --- Resilience's Cyber Mid Year Risk Report for 2026 highlights that human error, rather than AI-specific attacks, is the primary driver of cyber losses. The report shows that 85.3% of losses in the first half of 2026 were due to individuals falling for fraudulent communications. Despite the emergence of autonomous AI threats like JADEPUFFER ransomware and OpenAI's model breach, these did not contribute to the losses in Resilience's portfolio. Known vulnerabilities and governance-related issues were also significant causes of claims, though extortion remains the costliest. Claims frequency increased, driven by wrongful data collection claims, while average claim severity decreased due to fewer high-value extortion demands.
The findings from Resilience's report could influence underwriting strategies in the London Market, particularly in assessing human error risks. Insurers may consider revising policy wordings to address the rising trend of losses due to fraudulent communications. The absence of AI-specific attack losses suggests that current AI-related exclusions or coverages might not need immediate adjustment, but underwriters should remain vigilant as the threat landscape evolves. The increase in claims frequency and the shift in claim severity could impact pricing models and reserve allocations for cyber insurance products.