Additional AI regulation only for high-risk applications

The German insurance sector is calling for restraint in the regulation of AI-based applications. The current regulatory framework should be applied to its full extent before adopting new regulations, argues the German Insurance Association (GDV) in a 10-point plan. Insurance is already highly regulated and therefore does not need more rules. There are already many provisions governing data and consumer protection and they do not need to be reintroduced under the guise of AI supervision as they are technology-neutral.

The position paper was prompted by EU plans to establish a uniform legal framework for the use of artificial intelligence. The EU Commission aims to present an initial draft of the plan at the beginning of next year. The definition of AI will be pivotal: only applications with self-learning components should be subject to the new regulation.

Include AI in stronger sectoral supervision

While there are numerous ways to apply AI in insurance, such as claims management or forecasting future losses, the GDV believes that extra AI regulations should be restricted to high-risk areas, where there is a danger of violating people's basic rights or safety regulations. The GDV thus explicitly opposes the position espoused by the EU Parliament, i.e. that insurers and banks qualify as high-risk.

At the same time, the German insurance sector insists on reinforcing sectoral supervision. Its companies are monitored at national and European levels (BaFin and EIOPA respectively). Both authorities have developed their expertise in digitisation and AI significantly in recent years. As a result, there is no need for an additional supervisory authority in the insurance sector.