In 2023 (and 2020) Rocco Roberto Cerchiara conducted surveys on the use of ML in reserving in Italy.
The survey found that there was enthusiasm for developing techniques for machine learning in reserving. Whilst there has been a large interest in the actuarial offices, management is more cautious.
The main findings were:
- Companies are at very different stages in their research, but overall we are still at an early stage
- The main barriers companies were finding were:
- Data Quality
- Lack of best practice
- Time and resource constraints, and skills.
There was less appetite within companies with Standard Formula and/or small/medium companies.
Insurance companies reported that they were at an early knowledge-gathering stage regarding the use of ML algorithms, adopted mainly for the optimization of internal processes and, only in limited cases, in the relations with policyholders.
Further details are available in this slide deck.