Artificial Intelligence Governance Principles: towards ethical and trustworthy artificial intelligence in the European insurance sector

Updated on 06/01/2022

Artificial Intelligence Governance Principles: towards ethical and trustworthy artificial intelligence in the European insurance sector #

A report from EIOPA´s Consultative Expert Group on Digital Ethics in insurance #

European Insuance and Occupational Pensions Authority #

“From the Executive summary 

  1. “The advent of new technologies such as artificial intelligence (hereinafter “AI”), cloud
    computing or the internet of things (hereinafter IoT), coupled with the increasing availability of data in today’s digital society and economy, are enabling opportunities for future
    growth and development in the insurance sector… “
  2. “… The benefits arising from AI in terms of prediction accuracy, automation, new products and services or cost reduction are remarkable.
    However, there are also growing concerns amongst stakeholders about the impact that
    the increasing adoption of AI could have on the financial inclusion of groups of protected
    classes or vulnerable consumers or on our society as a whole….”
  3. “…With regards to the use of AI in insurance pricing and underwriting, the report includes
    guidance on how to assess the appropriateness and necessity of rating factors, noting
    that that correlation does not imply causation. Insurance firms should also avoid certain
    types of price and claims optimisation practices such as those aiming to maximise consumer’s “willingness to pay” or “willingness to accept”…”
  4. “…Finally, the report is based on the state-of-the-art of AI at the time of its publication. The
    GDE acknowledges that AI is an evolving technology with an ever-increasing number of
    applications and where extensive research is on-going. This is, particularly, the case in the
    area of transparency and explainability, as well as in the area of active fairness seeking to
    develop fairness and non-discrimination metrics to assess the outcomes of AI systems.
    As these areas of application and research evolve, the recommendations included in this
    report may also need to be revised in due course.”