In the future, AI is expected to help insurers to predict risk with greater accuracy and to use enhanced foresight to rapidly deploy new products in response to emerging risks.
However, insurers face three challenges in maximising the benefit for both consumers and themselves in their use of AI:
• Firstly, when developing AI systems, insurers face restricted access to data from the public sector. For maximum societal benefit, such datasets should be available for free and in a machine-readable format. The non-personal data for some AI applications is also sometimes concentrated in the hands of a few entities, resulting in restricted or expensive access to data that could improve AI systems and better serve customers. This raises questions over how access to this data should be governed.
• Secondly, insurers would like to see a holistic approach to regulation. The EU legal framework already covers areas relevant to AI such as fundamental rights, privacy and data protection, as well as product safety and liability. This is then complemented by national regulatory frameworks. To support the development and uptake of AI, and to avoid unnecessary regulatory burdens, a horizontal, proportionate and principles-based AI regulatory framework is needed that builds on existing EU and national regulatory frameworks.
• Finally, there is currently a lack of effective collaboration between authorities with responsibilities in the field of digitalisation. There is a need for all national authorities, whether they are responsible for conduct of business, prudential regulation, competition or data protection, to work closely together and ensure consistency in applying the rules to further develop the digital single market.
Moreover, the insurance industry supports the deployment of ethical, trustworthy and human-centric AI via an appropriate, risk-based and proportionate regulatory framework. The scope of the framework should be targeted only at those AI applications with proven high risk and significant effects on the rights of individuals, as not all uses of AI pose significant risks or directly impact consumers.
In the context of financial services legislation, and insurance in particular, principles such as transparency, fairness, accountability and ethics are to some extent already addressed by rules on conduct of business and disclosure, while rules on advice apply whether the recommendation is provided by a human or by AI.
|