- Publication date
- 13 September 2023
Artificial Intelligence (AI) is expected to play a pivotal role in the digital transformation across all industries and society as a whole. In the insurance sector, AI combined with the internet of things is providing for a wide range of opportunities for future growth and development. As shown by EIOPA’s thematic review on the use of Big Data Analytics in motor and health insurance, already in 2018, 31% of the participating European insurance firms were using AI across the insurance value chain, and another 24% were at a “proof of concept” stage. This trend was confirmed in the report on AI governance principles from EIOPA’s Stakeholder Group on Digital Ethics in 2021. In a workshop on AI organised by EIOPA earlier this year, industry participants further emphasized the transformative role of AI and provided clear examples of AI applications in the insurance industry.
While AI is already being used throughout the insurance value chain, recent advancements, such as the rise of generative AI, have demonstrated that we are only seeing the beginning of AI’s potential for the sector. The widespread adoption of generative AI by insurance companies is still in the early stages; however, insurers are actively exploring its potential uses, such as providing consumer advice, guiding policyholders through claims procedures, and enhancing pricing and underwriting processes.
In this context, the proposal of the European Commission for a Regulation on artificial intelligence (the AI Act) aims to lay down a uniform legal framework for the development and use of AI in conformity with European Union values. The proposal is currently being finalised by the EU co-legislators in trialogues.
The AI Act has specific implications for the insurance sector. Firstly, it is ‘horizontal’ and aims to cover all relevant sectors at once. Its cross-sectoral nature can raise challenges with integrating its provisions and their supervision into each sector, notably those that are highly regulated and supervised, such as the insurance sector. Secondly, the AI Act envisages harmonised standards being developed by the European Standardisation Organisations and the provision of guidance and compliance tools to aid providers and users in meeting the requirements.Thirdly, following a risk-based approach, the vast majority of the requirements of the AI Act apply to AI systems identified as high risk.
In light of the specificities of the insurance sector, the development of standards and guidance to facilitate the implementation of the AI Act will play a key role in ensuring a seamless application and preventing potential frictions with the insurance legislative and supervisory framework. While, based on the positions from the European Parliament and the Council, it seems likely that certain use cases in life and health insurance will be considered as high-risk, there are a wide range of other potential use cases where certain transparency obligations and voluntary codes of conduct would apply. In relation to the latter, additional efforts may be needed to ensure coherence between sectoral requirements and standards under the AI Act, while maintaining proportionality. In addition, the AI Act is not fully exhaustive in relation to the regulation and supervision of AI and sectoral legislation addressing conduct and prudential objectives continues to apply also to AI when used in the insurance and occupational pensions sectors. Recognizing the potential significance of AI for the insurance sector, EIOPA stands ready to contribute to its regulation and supervision. This commitment includes participating in the governance framework that will be established at the EU level.
As we look ahead, it is evident from the emergence of generative AI that the field of AI is evolving rapidly. The AI Act recognises this and aims to be future proof, allowing for adaptation to upcoming developments. The specific measures intended to achieve this goal are yet to be determined, particularly concerning the governance system, where proper consultation of sectoral expertise is important. Furthermore, the need to adapt to market changes has already been shown, as the Council and the Parliament introduced requirements for generative/foundation models that were not initially anticipated in the original proposal by the Commission. The final extent and scope of the measures are not yet clear, but these will have real impacts on the insurance and pensions sectors.
In conclusion, while AI is already being incorporated into the insurance value chain, the rise of generative AI suggests that its full potential is yet to be realized. The AI Act aims to provide a framework that will help the development of AI in conformity with European Union values. The future impact of the AI Act in the insurance sector depends on the outcome of trialogues and will necessitate the development of guidance for its effective implementation.