For a management briefing, begin with the NAIC overview and NIST framework below. For an actuarial learning plan, start with the SOA primer and CAS resources. Implementation documentation is most useful once you have defined a task and its acceptance checks.
For senior managers
NAIC · Artificial Intelligence topic page
Start with the NAIC's overview of insurer AI governance and its Model Bulletin. Use the applicable state insurance department's requirements for decisions in that jurisdiction.
Read the NAIC AI overviewNIST AI Risk Management Framework + Generative AI Profile
Use this voluntary framework to organize AI risk ownership, assessment, measurement, and management. Its Generative AI Profile addresses risks specific to generative systems.
Read the NIST framework and profileNAIC Model Bulletin adoption map (PDF)
A starting point for tracking state action. Check the map's date and confirm scope, wording, and effective dates with each relevant state insurance department.
Open the state adoption map (PDF)NAIC Big Data and Artificial Intelligence (H) Working Group
Meeting materials, pilot updates, and draft evaluation tools. Distinguish consultation drafts and pilot materials from requirements adopted by a state.
Review current working-group materialsQuarles & Brady · state adoption summary
Historical legal commentary on state adoption. Use it for background; use current NAIC and state publications to establish today's position.
Read the historical adoption commentaryVendor trust documentation
Use alongside the proposed contract to review security documentation, data use, retention, and access controls. Confirm terms for the specific product and account.
trust.anthropic.com · trust.openai.comFor actuaries: structured learning
SOA Research Institute · "A Primer on Generative AI for Actuaries" (2024)
An introduction to generative AI with actuarial examples. A useful first read before moving into implementation and validation.
Read the actuarial primerSOA Research Institute · "Operationalizing LLMs" (2025)
Follow the research index to the report on deploying and evaluating LLMs in actuarial work, then check the publication date before applying product-specific examples.
Browse the SOA AI research indexCAS · AI Tools and Resources hub
Property and casualty research and learning resources, including the AI Fast Track program. Use this alongside the SOA materials for practice-specific examples.
Browse CAS learning resourcesSOA PD recording · "Generative AI for Actuaries" (Jan 2025)
A recorded introduction for actuaries. Use it as a team discussion starter and check the publisher's access conditions.
Open the SOA recordingDeepLearning.AI short courses
Browse practical courses on prompting, retrieval, and building AI applications. Check each course's prerequisites, duration, and current pricing before enrolling.
Browse the course catalogFor actuaries: hands-on reference
Anthropic documentation & prompt-engineering guide
Primary implementation guidance for Claude. Treat its examples as provider-specific and evaluate them on your own task.
Read Claude documentationOpenAI Cookbook
Worked code examples for structured extraction, RAG, and evals.
Browse OpenAI worked examplesActuarial Standards Board · applicable standards
Read the standards themselves when assessing modeling, data quality, and communication responsibilities. Applicability depends on the assignment and the actuary's responsibilities.
ASOP 56: Modeling · ASOP 23: Data Quality · ASOP 41: Actuarial CommunicationsStanding intelligence: assign an owner, check quarterly
| What to watch | What the owner should record |
|---|---|
| NAIC Big Data & AI Working Group outputs | Track draft changes, pilot findings, and any state adoption; do not assume a pilot becomes a binding examination standard |
| Domiciliary and top-premium states' DOI bulletins | Record the requirements that apply to your legal entity, products, and states, including local variations |
| Colorado DOI algorithm/data-governance rulemaking | Track the rules and testing requirements for the lines of business within scope |
| SOA / CAS AI research index pages | Review new research for relevant methods, limitations, and evidence |
| Model provider release notes | Review version changes and deprecations, then re-run the evaluation set before adopting an update |