It will come as no surprise to many that the insurance industry is evolving with the opportunities that are being presented by artificial intelligence (AI). Investment in insurtech is at a record high, and a recent survey by Accenture showed that 84 per cent of insurers believe that AI will either ‘significantly change or completely transform’ the industry in the next three years.
While still at a nascent stage, technologies such as chatbots and pricing algorithms, combined with real-time sensors and advisory services that can help customers improve the quality of their lives, promise to bring big changes for customers and businesses alike across the value chain.
At a time of such significant change, it is inevitable questions of how to manage the risks emerge. Concerns about using data to influence customers, hyper-personalised insurance and social exclusion are real and will need a thoughtful approach to stay on the right side of the emerging regulation.
This month, we have had another hint that regulation of the insurance industry's use of AI is likely to be a priority in the form of a snapshot paper (PDF) from the Centre for Data Ethics and Innovation. The document provides helpful insight into the issues the industry is facing and suggests ways in which both the industry and government could be moving forward.
This area is complex and fast evolving. So, to get you started, here are three things you should be doing now and three things you should be considering.
You should be:
- reviewing data-collection strategies, particularly those involving new sources of data such as wearables, telematics and social media;
- establishing governance frameworks and reviewing expertise to deal with the specific issues being raised by AI; and
- making privacy notices more accessible, considering innovative techniques and making use of technologies such as haptic feedback or pressure sensitive displays.
You should be considering:
- carrying out data discrimination audits to check algorithms and training datasets for unwarranted bias;
- using public engagement to test the market on difficult issues such as transparency; and
- sharing customer data within the industry (data pooling) to improve ‘switching’ for customers.
