Log in / My SOA

Model Governance/Ethics for GenAI Models

Background and Purpose

Recent regulatory developments around the ethics of using AI based models attracted lots of publications and industry debates. Generative AI (GenAI) poses unique challenges, such as the complexity, hallucination, reproducibility, etc. During the model development stage and thru the applications, GenAI can have profound impacts to the society and life insurance industry. Actuaries already started building own GenAI-based models. It is both timely and critical to have a set of guidelines to assist those actuaries and provide an educational piece to actuaries who are interested in the topic.

The Actuarial Innovation and Technology Steering Committee (AITSC) acknowledges that there are no universally accepted definitions of ethics. While papers and frameworks exist, this project aims to bridge the gap between theory and practice by providing actuaries and data scientists with actionable tools, templates, and processes to productize GenAI models responsibly.

The research should address the real-world challenges actuaries face when integrating AI models into their workflows, helping them overcome ambiguity and subjectivity. The deliverables should include step-by-step processes, templates, and checklists to ensure models are ethically sound, compliant with regulations, and effective in actuarial applications.

The AITSC encourages researchers to respond to this RFP even if their bid does not perfectly match the purpose above. The AITSC is eager to collaborate with researchers on a mutually agreed scope of research.

Scope of Work

The proposal should deliver clear, actionable resources in the following areas:

  1. Model Governance Framework
    • Step-by-Step Model Validation: Provide a practical process for validating GenAI models before deployment. This should include specific tests, metrics, and benchmarks for accuracy, fairness, hallucination, reproducibility, and reliability in actuarial tasks.
    • Continuous Monitoring and Reporting: Outline an actionable process for monitoring GenAI models post-deployment. Include guidelines for performance reviews, recalibration, and troubleshooting, along with templates for regular reporting to stakeholders.
    • Audit Framework: Develop an easy-to-use audit framework that includes templates for documenting model decisions, assumptions, and any updates made during the model’s lifecycle. The goal is to facilitate model transparency for internal or external audits.
  2. Ethical Considerations
    • Bias Mitigation Action Plan: Provide a clear, actionable methodology for detecting and mitigating bias in training data and model outputs. This should include practical tools (e.g., bias detection algorithms, example workflows) and how to incorporate fairness checks.
    • Accountability Framework: Propose a concrete framework that defines how responsibility for model decisions is tracked and attributed. This should include a practical way to record decision rationale and ensure appropriate oversight, along with accountability checkpoints throughout the model’s lifecycle.
    • Transparency Guidelines: Recommend simple, effective ways to make GenAI model predictions interpretable for actuaries and stakeholders. The solution should include tools for ensuring that model decisions can be explained and audited without requiring advanced AI expertise.
  3. Regulatory Compliance
    • Compliance Checklist: Develop a checklist for ensuring that GenAI models follow life insurance regulations (e.g., NAIC Model Bulletin, New York DFS circular letter, Colorado SB21-169). The checklist should highlight regulatory requirements specific to GenAI used in actuarial contexts and provide actionable steps for meeting each requirement.
    • Documentation Standards: Provide a standardized framework for documenting the model development process to demonstrate compliance with relevant regulations. This framework should include specific fields and templates for model documentation, such as how to report training data sources, model assumptions, and evaluation results.
  4. Best Practices for Productization
    • Integration Plan: Provide a practical integration plan for embedding GenAI models into existing actuarial workflows (e.g., pricing, underwriting, valuation, claims management). The plan should include step-by-step instructions for ensuring smooth integration, from testing to deployment, with specific action items.
    • User Training and Support Materials: Develop training materials for actuaries that explain how to use GenAI models in practice. The materials should suit non-technical users and include practical scenarios and real-world examples of how GenAI models can be applied to actuarial tasks.
    • Ethical Productization Guidelines: Offer clear, actionable guidelines for ensuring that ethical principles are embedded throughout the productization process. This includes addressing issues such as model interpretability, explainability, fairness, and the responsible use of customer data.

Note that the list above is not meant to be exhaustive but merely examples of proposed topics that may be researched.

Proposal Requirements

To facilitate the evaluation of proposals, the following information should be submitted:

  1. Resumes of the researcher(s), including any graduate student(s) expected to participate, indicating how their background, education and experience bear on their qualifications to undertake the research. If more than one researcher is involved, a single individual should be designated as the lead researcher and primary contact. The person submitting the proposal must be authorized to speak on behalf of all the researchers as well as for the firm or institution on whose behalf the proposal is submitted.
  2. An outline of the approach to be used (e.g. literature search, model, etc.), emphasizing issues that require special consideration. Details should be given regarding the techniques to be used, collateral material to be consulted, and limitations of the analysis.
  3. A description of the expected deliverables and any supporting data, tools or other resources.
  4. Cost estimates for the research, including computer time, salaries, report preparation, material costs, etc. Such estimates can be in the form of hourly rates, but in such cases, time estimates should also be included. Any guarantees as to total cost should be given and will be considered in the evaluation of the proposal. While cost will be a factor in the evaluation of the proposal, it will not necessarily be the decisive factor.


    Please note that as a policy, the SOA Research Institute generally does not provide funding to cover academic institution overhead expenses.

    As a guide for developing the project budget, please review the Historical Project Cost Guide (see Appendix)

  5. A schedule for completion of the research, identifying key dates or time frames for research completion and report submissions. The AITSC is interested in completing this project in a timely manner. Suggestions in the proposal for ensuring timely delivery, such as fee adjustments, are encouraged.
  6. Other related factors that give evidence of a proposer's capabilities to perform in a superior fashion should be detailed.

Selection Process

The AITSC will appoint a Project Oversight Group (POG) to oversee the project.  The [sponsoring entity/POG] is responsible for recommending the proposal to be funded. Input from other knowledgeable individuals also may be sought, but the AITSC will make the final recommendation, subject to Society of Actuaries Research Institute (SOA) leadership approval. An SOA staff research actuary will provide staff actuarial support.

Questions

Any questions regarding this RFP should be directed to Research-AIT@soa.org.

Notification of Intent to submit Proposal

If you intend to submit a proposal, please email written notification by May 2, 2025 to Research-AIT@soa.org.

Submission of Proposal

Please email your proposal to Research-AIT@soa.org; proposals must be received no later than May 30, 2025. It is anticipated that all proposers will be informed of the status of their proposal by the end of June 2025.

Conditions

The selection of a proposal is conditioned upon and not considered final until a Letter of Agreement is executed by both the Society of Actuaries Research Institute and the researcher.

The AITSC reserves the right to not award a contract for this research. Reasons for not awarding a contract could include, but are not limited to, a lack of acceptable proposals or a finding that insufficient funds are available. The AITSC also reserves the right to redirect the project as is deemed advisable.

The AITSC plans to hold the copyright to the research and to publish the results with appropriate credit given to the researcher(s).

The Society of Actuaries Research Institute and AITSC may choose to seek public exposure or media attention for the research. By submitting a proposal, you agree to cooperate with the [Society of Actuaries/sponsoring entity] in publicizing or promoting the research and responding to media requests.

The Society of Actuaries and AITSC may also choose to market and promote the research to members, candidates and other interested parties. You agree to perform promotional communication requested by the Society of Actuaries Research Institute and AITSC, which may include, but is not limited to, leading a webcast on the research, presenting the research at an SOA meeting, and/or writing an article on the research for an SOA newsletter.

Conflict of Interest

You agree to disclose any of your material business, financial and organizational interests and affiliations which are or may be construed to be reasonably related to the interest, activities and programs of the Society of Actuaries Research Institute and AITSC.

Appendix

The cost ranges below are intended as a guide for budgeting project costs for proposals in response to SOA Research Institute Request for Proposals (RFP). Please note these figures span the 33rd to 66th percentiles for all projects as well as projects that involve a specific approach (lit review, survey, etc.). They are based on historical costs over several recent years. Expected costs for some RFPs may fall outside these ranges depending on the nature of the work and resources required for completion.

All Contracted Projects

This category includes all contracted projects that the Institute has undertaken within the last several years.

The 33rd-66th percentile project costs range is $25,000 - $50,000.

Literature Reviews

This category includes projects that involved only a literature review or the cost for the portion of a larger project that included a literature review. 

The 33rd-66th percentile project costs range is $15,000 - $20,000.

Surveys

This category includes all projects that had a survey as their primary component. 

The 33rd-66th percentile project costs range is $28,000 - $55,000.