The European Insurance and Occupational Pensions Authority (EIOPA) launched a public consultation focused on AI governance and risk management. The consultation will shape supervisory guidance on how to apply existing insurance provisions to AI systems.
This move increases regulatory pressure and compliance costs for the insurance sector, Fitch Ratings says.
EIOPA’s Digitalisation Report from May 2024 shows growing AI use. About 50% of non-life insurers and 24% of life insurers currently apply AI in operations.
Adoption is expected to rise sharply due to generative AI and foundation models.
AI already supports pricing, underwriting, claims handling, fraud detection and sales. However, its wider use creates new risks. AI systems lack transparency, making decision processes hard to explain.
Errors in training data can cause biased outcomes. Dependence on large datasets and algorithms increases the chance of pricing mistakes and other liabilities.
In response, the European Commission’s AI Act, introduced in July 2024, sets clear risk management requirements. It applies a four-tier system—minimal, limited, high and unacceptable risk—with varying obligations.
High-risk tools, such as those in life and health risk assessments, must meet strict standards. Other AI uses remain subject to current insurance rules.
The Act’s cross-sector application may cause issues when aligning it with insurance-specific rules. Insurers may need to expand internal capacity or collaborate with external providers to manage compliance effectively.
Most of the AI Act takes effect in August 2026, giving insurers time to prepare. Still, failure to comply could lead to heavy penalties.
Firms must also adjust existing systems to meet new standards, which will demand further spending. While tighter rules bring extra costs, they may help maintain sound practices across the sector.
