Call for Participation: AI & Individual Life Insurance Challenge
Issued by the Mortality & Longevity Strategic Research Program
Background and Purpose
The SOA Research Institute’s Mortality & Longevity Strategic Research Program (MLSRP) invites participation in the AI & Individual Life Insurance Challenge, an initiative designed to explore how artificial intelligence (AI) can be applied to individual life insurance data. AI methods, including large language models, machine learning, and deep learning, are increasingly relevant to actuarial practice, offering new ways to analyze experience data, detect emerging patterns, and improve forecasting.
This challenge focuses on applying AI approaches to a publicly available individual life insurance dataset. Participants are asked to evaluate the performance, interpretability, and practical utility of their chosen methods while also providing actionable insights for practitioners. The goal is to demonstrate how AI can complement traditional actuarial techniques, enhance understanding of experience data, and contribute to better decision-making across pricing, valuation, and risk management.
Data and Prior Analysis
The data to be used for this analysis and related existing reports can be found:
2019 Individual Life Insurance Mortality Experience Report | SOA
Note that the other material on this webpage related to copula models and tableau dashboards are not required of this project.
Research Objectives
Submissions should balance attention to methodological approach and practical insights. Specifically, participants are expected to:
- Justify the choice of AI techniques, including considerations of model design, transparency, interpretability, and appropriateness for actuarial practice.
- Demonstrate what practitioners can learn from the application of these methods, highlighting strengths and limitations compared to traditional actuarial techniques.
- Communicate insights drawn from the dataset, including trends, anomalies, or actionable findings that could inform real-world actuarial decision-making.
- Present results in a way that is understandable to both technical and non-technical stakeholders.
In addition, participants are encouraged to consider the use of LLM tools (e.g. ChatGPT) to accomplish some of the work (could be idea generation, research, writing code, writing the paper, data analysis), and explain how they were used. An important goal of the research is to educate actuaries on AI methods, not limited to the more advanced models.
Deliverables
Each submission must include:
- A written report outlining data preparation, methodological approach, results, and discussion of implications.
- Exhibits or visualizations that support findings and aid communication.
- A 1–2 page summary of key practitioner takeaways.
Supporting code or documentation sufficient for reproducibility.
Evaluation Criteria
Entries will be evaluated based on the following criteria:
- Advancement of Knowledge – Creativity, originality, and contribution to actuarial practice, with emphasis on educating actuaries about AI methods, including but not limited to advanced models.
- Approach – Appropriateness, innovation, and justification of AI methods.
- Clarity – Quality of explanation, organization, and communication.
- Insights – Relevance and usefulness of findings for practitioners and the life insurance industry.
- Practicality – Consideration of implementation, interpretability, and alignment with actuarial standards.
Copyright Notice
Award recipients will be required to assign all copyrights in their work to the SOA Research Institute.
Eligibility
The challenge is open to all, including but not limited to academic researchers, industry practitioners, AI tech experts, students, and interdisciplinary teams. Collaboration across backgrounds is encouraged.
Timeline
The following dates are anticipated but subject to adjustment:
- Challenge announcement and dataset release: early October
- Submission deadline: January 31
- Review and selection of finalists: early April
Awards
The SOA Research Institute has allocated $22,500 in award funding for this challenge. Awards will be distributed as follows:
- First Place: $10,000
- Second Place: $7,500
- Third Place: $5,000
Publication and Presentation
Selected submissions will be published electronically by the SOA Research Institute. Winning teams may also be invited to present their findings at an SOA meeting, webcast, or other professional development event. The Institute may seek additional venues to broaden dissemination of the work.
Rules for Submission
- Submissions must be in English. Written reports should be 3,000–6,000 words.
- Each submission must include author(s) name(s), credentials/designations (if applicable), title, organization, and contact information.
- Content must not contain overt political statements, commercial promotion, or other inappropriate material.
- Submissions must comply with SOA antitrust guidelines.
- Submissions should be sent electronically to research-ML@soa.org. When submitting, please use subject line “ML281 - Call for Participation: AI & Individual Life Insurance Challenge”.
Rights Granted
By submitting, authors grant the SOA Research Institute an unlimited license to publish, republish, or otherwise disseminate their submission with proper attribution.
Questions
Questions regarding this challenge may be directed to research-ML@soa.org.