The Challenges of AI to Insurability

Insurability depends heavily on understanding well the risks of those to be included in the insured pool. Without good data and advance modelling it will be impossible to quantify the level of risk and the correct pricing.
There have been concerns that new opportunities, i.e., increased amounts of data, more developed predictive methods and use of artificial intelligence or machine learning can present challenges to the earlier concept of insurability. Such concerns could be:

• It can become possible to classify risks so minutely that the earlier risk-sharing becomes obsolete, or
• Increased amounts of tools may tilt the information asymmetry (traditionally benefiting the insured) to its top (i.e. the insurer might have tools that make them far better in assessing the risk than the insurer).

Prior research suggests that the impact of artificial intelligence on society has so far been largely positive, reshaping the way financial institutions work and helping them in designing new products within the market. The Artificial Intelligence & Data Science Working Group of the Actuarial Association of Europe is preparing a paper on these topics, with the presenter leading this project. The paper intends to shed light on the topic by explaining insurability, taking a look at the state-of-the-art tools and analysing the tensions between insurability and novel data-related technologies.
 
Find the Q&A here: 'The Challenges of AI in Society'