The paper explores the use of machine learning, particularly deep learning techniques, in insurance pricing by modeling claim frequency and severity data. It compares the performance of various models, including generalized linear models and neural networks, on insurance datasets with diverse input features. The authors use autoencoders to process categorical variables and create surrogate models for neural networks to translate insights into practical tariff tables.
“... we provide evidence that setting the risk-weighted ratio at 15\% and the simple leverage ratio at 10% would significantly decrease the probability of default without hampering banks' activities.”
The study investigated how ChatGPT and Bard categorize investor risk profiles compared to financial advisors. While there were no significant differences in the risk scores assigned by the chatbots an
The paper suggests that companies developing high-risk AI systems should demonstrate their safety before deployment, arguing for proactive risk management. It proposes a risk management approach where