Explainable Artificial Intelligence (XAI) in Actuarial Pricing and Reserving: A Systematic Literature Review of Methods, Regulatory Compliance, and Predictive Trade-offs
Keywords:
Explainable Artificial Intelligence (XAI); actuarial science; machine learning; insurance regulation; risk classification.Abstract
The adoption of complex machine learning models in actuarial pricing and reserving has created a fundamental tension
between predictive accuracy and regulatory transparency. In this systematic review, we evaluate how Explainable Artificial
Intelligence (XAI) methods are being incorporated into the actuarial workflow to address this challenge. We conducted a
systematic search across Scopus, Web of Science, and IEEE Xplore for peer-reviewed papers published between January
2018 and June 2024, synthesizing 15 core studies based on the PRISMA 2020 guidelines. The results demonstrate that SHAP
(SHapley Additive exPlanations), and notably TreeSHAP, dominates the literature (representing 40% of studies), with
primary applications concentrated in motor pricing and underwriting rather than claims reserving. Rather than aggregating
mathematically non-comparable indicators, our synthesis disaggregates results by metric family, showing that post-hoc
explainer layers induce tight, minor performance penalties across likelihood, distance-based, and threshold metrics relative
to unconstrained black-box baselines, while consistently outperforming traditional Generalized Linear Models (GLMs).
However, critical gaps remain: no reviewed study conducted formal legal or empirical user validation, and assumptions
regarding compliance with frameworks like the GDPR or the EU AI Act are often discursively asserted rather than tested.
We conclude that while XAI techniques are rapidly propagating through actuarial research, transitioning these frameworks
into regulated practice requires a shift from passive post-hoc approximations toward rigorous causal modeling and
formalized legislative sandbox testing.
References
Arrieta, A. B., Díaz-Rodríguez, N., Del Ser, J., Bennetot, A., Tabik, S., Barbado, A., García, S., Gil-López, S., Molina, D., Benjamins, R., Chatila, R., & Herrera, F. (2020). Explainable Artificial Intelligence (XAI): Concepts, taxonomies, opportunities and challenges toward responsible AI. Information Fusion, 58, 82–115. https://doi.org/10.1016/j.inffus.2019.12.012
Azzone, M., Viarengo, P., & Andreatta, G. (2023). A machine learning model for lapse prediction in life insurance contracts. Expert Systems with Applications, 223, 116261. https://doi.org/10.1016/j.eswa.2021.116261
Baudry, M., Martin, C., & Pesenti, S. M. (2025). A machine learning approach for individual claims reserving in insurance. Applied Stochastic Models in Business and Industry. https://doi.org/10.1002/asmb.2442
Blier-Wong, C., Cossette, H., Lamontagne, L., & Marceau, É. (2021). Machine learning in insurance pricing: A review of interpretability methods with an application to cyber risk. North American Actuarial Journal, 25(4), 531–554. https://doi.org/10.1080/10920277.2021.1923055
Delong, Ł., Kozak, A., & Szymański, P. (2021). Collective attention networks for telematics car insurance pricing. ASTIN Bulletin: The Journal of the IAA, 51(2), 511–543. https://doi.org/10.1017/asb.2021.21
Denuit, M., Hainaut, D., & Trufin, J. (2020). Effective additive models for actuarial risk classification. ASTIN Bulletin: The Journal of the IAA, 50(2), 547–580. https://doi.org/10.1017/S174849952000005X
Embrechts, P., & Wüthrich, M. V. (2022). Recent challenges in actuarial science. Annual Review of Statistics and Its Application, 9, 119–140. https://doi.org/10.1146/annurev-statistics-040120-030244
Gabrielli, A. (2020). Neural network embedding of individual claims reserving models with explainable extensions. Scandinavian Actuarial Journal, 2020(8), 696–725. https://doi.org/10.1080/03461238.2020.1758414
Gan, G., & Valdez, E. A. (2020). Valuation of large portfolios of variable annuities using machine learning surrogate metamodels. ASTIN Bulletin: The Journal of the IAA, 50(1), 209–236. https://doi.org/10.1017/asb.2020.8
Henckaerts, R., Côté, M.-P., Antonio, K., & Verbelen, R. (2021). Boosting insights in insurance pricing: Generalized additive models versus XGBoost through an explainable lens. North American Actuarial Journal, 25(4), 555–577. https://doi.org/10.1080/10920277.2021.1934783
Kuo, K. (2019). DeepTriangle: A deep learning approach to loss reserving. Risks, 7(3), Article 97. https://doi.org/10.3390/risks7030097
Lorentzen, C., Mayer, M., & Meier, N. (2022). Peeking into the black box: An actuarial case study for interpretability of machine learning models. ASTIN Bulletin: The Journal of the IAA, 52(1), 97–131. https://doi.org/10.2139/ssrn.4110398
Lundberg, S. M., & Lee, S.-I. (2017). A unified approach to interpreting model predictions. Advances in Neural Information Processing Systems, 30, 4765–4774.
Nikolaos, A., Lekka, D., & Vlachos, D. (2024). Multimodal explainable artificial intelligence: A comprehensive review of methodological advances and future research directions. IEEE Access, 12. https://doi.org/10.1109/ACCESS.2024.3467062
Page, M. J., McKenzie, J. E., Bossuyt, P. M., Boutron, I., Hoffmann, T. C., Mulrow, C. D., Shamseer, L., Tetzlaff, J. M., Akl, E. A., Brennan, S. E., Chou, R., Glanville, J., Grimshaw, J. M., Hróbjartsson, A., Lalu, M. M., Li, T., Loder, E. W., Mayo-Wilson, E., McDonald, S., … Moher, D. (2021). The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. BMJ, 372, Article n71. https://doi.org/10.1136/bmj.n71
Richman, R. (2020). AI in actuarial science—A review of recent advances: Part 2. Annals of Actuarial Science, 15(2), 230–258. https://doi.org/10.1017/S1748499520000048
Spencer, K., Frees, E. W., & Valdez, E. A. (2022). mSHAP: SHAP values for two-part models in longitudinal insurance analytics. Risks, 10(1), Article 3. https://doi.org/10.3390/risks10010003
Wu, X., Meng, S., & Chan, F. T. S. (2022). Interpretable machine learning for health underwriting: A LIME-based application on electronic health records. Diagnostics, 13(16), 2681. https://doi.org/10.3390/diagnostics13162681
Wüthrich, M. V., & Merz, M. (2023). Statistical foundations of actuarial learning and its applications. Springer Actuarial. https://doi.org/10.1007/978-3-031-15211-5
