Skip to sources
Time Machine

Actuaries

Scrub through 274years of this role's history, from when it first emerged, through every wave of technology that reshaped it, to the cited projections for where it's heading next.

2026drag to travel through time
1775180018251850187519001925195019752000now
2026
Known today as Actuary — FSA / FCAS / ASA / ACAS (Society of Actuaries 1949; dual-credential structure stable through present)
US Employment
27K
OEWS is a point-in-time survey snapshot, not a continuous time series; BLS advises against using it for year-over-year trend comparison.
Median Annual Wage
$130,000
≈ $126,667 in 2024 dollars
Each dot is a cited figure over time; the dotted line only links them (values between aren't measured). Hollow dots are estimates.
Tools of the era

The tools that defined the work

Select an era to see how it reshaped the work.

  • Mortality tables, logarithm tables, and hand calculation

    Edmond Halley's 1693 Breslau life table (Philosophical Transactions of the Royal Society, No. 196) gave actuaries their first rigorous empirical foundation for pricing life annuities: a tabulated count of survivors at each age from a city-wide birth cohort. For the next two centuries, actuarial computation meant manually constructing and applying life tables using logarithm tables and arithmetic shortcuts called 'commutation functions' — algebraic condensations that reduced compound-interest-and-mortality calculations to lookups in precomputed columns. Abraham de Moivre's 1725 'Annuities upon Lives' introduced a simplified uniform-distribution approximation for mortality, giving actuaries a practical tool before full life tables were available. Every premium calculation was carried out by hand, with errors checked by a second actuary performing the same arithmetic. The job was essentially the work of a human computer.

    Work toolChanging equipment
  • Mechanical calculators (Marchant, Comptometer, Burroughs)

    Mechanical adding and multiplying machines — the Burroughs adding machine (1888), the Comptometer (1887), and the Marchant calculator (adopted widely by the 1920s) — accelerated actuarial arithmetic without displacing the actuary's judgment function. Insurance company "computation departments" staffed largely by women performed the mechanical steps under an actuary's direction. The principal computational product was the actuarial reserve calculation — the present value of future benefit payments — which required thousands of multiplications per policy for large portfolios. Slide rules were standard individual tools. The 1941 American Experience Table of Mortality (standard in US life insurance pricing through the 1950s) was constructed using these mechanical tools on data from nineteen large life insurers.

    Mechanical calculationTen-key speed
  • IBM mainframes and FORTRAN — reserve and valuation routines

    Travelers Insurance Company installed one of the first IBM mainframe computers in the 1960s; the IBM System/360, introduced in 1964, became the industry workhorse for actuarial valuation. FORTRAN — released by IBM in 1957 — became the programming language of choice for actuarial reserve models. For the first time, a life insurer could mechanically value every policy in force overnight rather than sampling or approximating. The mainframe did not eliminate actuarial judgment; it redirected it from arithmetic to model design. The job of the valuation actuary shifted from calculating reserves manually to writing, validating, and governing the programs that calculated them. Mainframe computing was a shared resource — batch jobs queued overnight — which constrained the actuary's ability to run scenarios interactively.

    Effect on the work

    Mainframe computing dramatically reduced the clerical actuarial staff ("computation departments") of insurance companies while increasing demand for actuaries who could specify and validate the machine models. The net effect on the credentialed actuarial profession was employment growth, not displacement.

    Mainframe processingComputerized records
  • MoSes and domain-specific actuarial modelling languages (1970s-80s)

    MoSes (Model of Stochastic Events System), developed by Tillinghast (later Towers Perrin, now Moody's Analytics), emerged in the late 1970s and early 1980s as the first domain-specific actuarial modelling language — purpose-built for life insurance product pricing, valuation, and profit testing. Unlike FORTRAN, which required programming expertise, MoSes allowed actuaries to describe policy structures and projection logic in their own vocabulary. It became the standard platform for complex life product modelling at large insurers and consulting firms through the 1980s. APL (A Programming Language), an array-oriented language popular in IBM environments, was also widely adopted by insurance companies in this era for its ability to manipulate actuarial tables as arrays without nested loops.

    Work toolChanging equipment
  • Lotus 1-2-3 and PC spreadsheets — actuarial liberation from the mainframe

    The early 1980s PC revolution liberated actuaries from the mainframe's overnight batch queue. Lotus 1-2-3 (1983) — running on the IBM PC — gave individual actuaries interactive computation for the first time: build a mortality table, change an assumption, see the result in seconds. The effect on actuarial practice was profound. 'The personal computer was liberating — the actuary was free,' as SOA Systems Evolution histories describe it. Actuaries could now prototype pricing models, run sensitivity analyses, and present results to management in a single day, rather than waiting for the data-processing department. The spreadsheet also democratized actuarial modelling within insurance companies: finance and underwriting staff could interrogate actuarial results without going through the actuarial department.

    Mainframe processingComputerized records
  • Prophet and AXIS — enterprise actuarial valuation platforms

    Prophet (originally developed in the early 1990s; later acquired by SunGard, then FIS in 2015) and AXIS (developed by Milliman's actuarial consulting practice) became the dominant actuarial cash-flow projection platforms for life insurance valuation through the 1990s and 2000s. Unlike spreadsheets, these platforms could project the full liability of an insurer's in-force book — hundreds of thousands of policies across multiple product lines — into hundreds of future time periods, applying stochastic interest and mortality scenarios. They produced the reserve exhibits required by NAIC statutory accounting, GAAP, and (from 2023) IFRS 17. 'Prophet is a projection system and can be compared to an Excel spreadsheet on steroids,' as one SOA practitioner described it in 2013. Prophet served 10,000+ users in nearly 1,000 sites across 70+ countries at its peak.

    Accounting softwareIntegrated ledgers
  • R and Python — predictive modeling enters actuarial practice

    The shift from traditional GLMs constructed in Excel to gradient-boosting models, random forests, and neural networks deployed in R and Python redefined what actuarial pricing models could do. By the mid-2010s, personal auto insurers were building telematics-based pricing models using sensor data; Lemonade (founded 2015, launched NY 2016) built its entire underwriting and claims operation on machine learning from day one. Usage-based insurance — pricing based on actual miles driven, time of day, and driving behavior — required ML infrastructure that spreadsheets and Prophet could not deliver. By 2020, Python had become a listed job-posting requirement at major US carriers (per SOA research), and the CAS launched its Exam MAS series explicitly covering predictive analytics and machine learning.

    Effect on the work

    Python/ML skills widened the gap between junior actuaries proficient only in Excel-and-Prophet and those who could build end-to-end ML pricing pipelines. Early evidence (hyperexponential 2025 research) showed that data-manipulation work consumed roughly 70% of a junior actuary's time before ML platforms arrived; with ML tooling, that dropped to ~30% — compressing the entry-level work pyramid.

    Work toolChanging equipment
  • Akur8, hyperexponential (hx), WTW Radar 5 — AI-native actuarial pricing platforms

    Akur8 (founded 2019; expanded to 300+ carrier clients in 40+ countries by 2025) and hyperexponential hx Renew brought transparent machine learning — gradient-boosting models with actuarially interpretable outputs — directly into the pricing actuary's workflow, replacing the weeks-long cycle of building GLMs in Excel with an ML pipeline that delivers model outputs in hours. Akur8 expanded its platform further by acquiring Arius (Milliman's reserving software) in September 2024 and Slope Software (life actuarial modeling) in 2024, building an end-to-end actuarial AI platform spanning pricing, reserving, and life modeling. WTW Radar 5, relaunched with generative AI capabilities in October 2025, added natural-language model construction. hyperexponential's own research documented that AI deployment cut actuarial data-manipulation time from 70% to 30% of total work hours.

    Effect on the work

    Entry-level actuarial work — chain-ladder runs, factor triangulations, experience-rating spreadsheets — is the first category being automated by these platforms, compressing the base of the traditional junior-actuary work pyramid. The net effect on total actuary headcount is contested: BLS projects +22% growth despite (or because of) AI adoption, as carriers need more actuaries to govern the AI models than to run the manual calculations they replace.

    Work toolChanging equipment
  • Generative AI for actuarial memos, IFRS 17 reporting, and reserve narratives

    The arrival of GPT-4 in March 2023 and Claude 3 in 2024 introduced LLMs into the writing-intensive portions of actuarial work: Statement of Actuarial Opinion narratives, actuarial memoranda supporting rate filings, DOI interrogatory responses, and the new IFRS 17 disclosure requirements (effective 1 January 2023) that significantly expanded the volume of written actuarial narrative required per reporting period. The SOA's January 2026 newsletter "Navigating the AI Transformation in Actuarial Science" described governance, oversight, and accountability decisions as the profession's strongest AI moat — not because LLMs cannot draft the prose, but because the actuary's personal professional attestation (SAO, rate-filing certification, ERISA pension stamp) cannot be delegated to any model regardless of its prose quality. CAS launched an AI Fast Track Program in 2025, acknowledging that ML-based stochastic reserving reduces analysis time by 50%.

    AI audit toolsPattern detection
Projection cone · present → 2034

What credible sources project

Scrub the slider past now to anchor each scenario on the scrubber. The spread is the range of futures credible sources project for this role.

Employment outlook
Projected change in the number of people doing this work.
BLS Occupational Outlook 2024-34
2034
+22%
BLS Employment Projections — industry-occupation matrix + labor productivity assumptions + replacement-need modeling. The 2024-34 OOH projects 22% employment growth for actuaries — roughly 4× the 5% national average — driven by growing demand for AI model governance, expanding health insurance markets, enterprise risk management, and climate-related catastrophe modeling. Unlike the flat projection for paralegals (where BLS explicitly cites AI as a constraint), the actuary OOH notes that AI creates actuarial demand by generating more models that require oversight and validation. ~2,400 annual openings are projected from growth plus turnover replacement. The BLS projection uses 33,600 as the 2024 baseline.
Deloitte — 2025 Insurance Industry Outlook
2030
+10%
Deloitte's 2025 Insurance Industry Outlook survey found that 82% of insurance carriers plan AI adoption primarily to address labor shortages — not to reduce headcount — and that actuarial model governance roles are among the positions most actively being created by AI adoption, not displaced by it. The net employment effect of AI on actuaries under this framing is moderately positive: carriers need actuaries to certify AI-generated reserve and pricing outputs, not just to produce those outputs themselves. The +10% projection represents Deloitte's implicit demand expansion scenario if AI adoption expands the actuarial addressable market (more risk types modeled, more geographies, more granular pricing segmentation) faster than it automates existing tasks.
AI task exposure
Share of the role’s tasks that researchers estimate AI can do. This is a measure of task exposure, not a forecast of jobs lost.
Eloundou et al. — "GPTs are GPTs" (2023)
2030
30%
of tasks
GPT-4 task-by-task labeling against O*NET task statements for 15-2011. Actuaries receive a materially higher exposure score under Eloundou's β metric than under Frey/Osborne's framework — because β measures how much an LLM could assist with tasks (capability), not whether the task can be fully automated. Actuarial tasks like "draft written reports" and "present findings to management" score high on β because LLMs can assist with the writing and structuring, even if the actuary's professional certification cannot be delegated. Finance and insurance professionals overall score β ≈ 0.35-0.45 in Eloundou (2023). The -30% figure here represents an upper-bound task-exposure interpretation for actuaries — not a projected employment loss — and should be read as "up to 30% of actuarial task time could be accelerated by LLMs" rather than "30% of actuarial jobs will be eliminated."
Frey & Osborne (2013)
2033
4%
of tasks
Gaussian-process classifier on O*NET task features; 702 occupations rated by computerisation probability. Actuaries scored among the lowest-risk occupations in the entire study — estimated computerisation probability approximately 0.04 (4%), placing actuaries in the "very low risk" category alongside surgeons, social workers, and dentists. The low score reflects the credential-anchored regulatory tasks (Appointed Actuary SAO sign-off, ERISA pension certification, rate filing attestation) and the "social intelligence" and "creative intelligence" bottlenecks that Frey/Osborne identified as automation barriers: actuaries regularly present model uncertainty to boards and regulators, tasks their framework rated as non-computerisable. Note: the F&O probability is for the core task set, not a direct employment-loss forecast; the -4% projection here represents a conservative interpretation of a low-automation occupation under their framework.
Today, in this role

What's shifting in the work right now

The historical view above shows how this role has moved. This is the present-day detail: which AI tools are picking up which tasks, where the edge still is, and the natural directions this work can grow.

What's changing in your day

Three parts of your work where AI is already doing real lifting, and what stays yours.

AI is sitting alongside you hereWrite Python or R code for bespoke actuarial analyses — experience studies, credibility-weighted development factors, frequency-severity separation, predictive loss cost models — using GitHub Copilot for code generation acceleration

Write Python or R code for bespoke actuarial analyses — experience studies, credibility-weighted development factors, frequency-severity separation, predictive loss cost models — using GitHub Copilot for code generation acceleration; review and validate AI-generated statistical code against actuarial standards before it is used in reserve or pricing deliverables that carry professional sign-off.[10],[11]

Where your edge is

AI code generation for actuarial work is fast but requires rigorous validation: Copilot hallucinations in chain-ladder factor selection or credibility-weighting syntax can produce plausible-looking but actuarially wrong outputs. Build a test-suite discipline — always validate generated code against a small dataset with known analytical results before applying it to production reserve or pricing work. Python proficiency with scikit-learn, pandas, and statsmodels is now listed in job postings by major carriers as a baseline requirement.

AI is sitting alongside you hereBuild and validate insurance pricing models using Akur8 or hyperexponential (hx Renew): use the platform's transparent ML engine to fit GLMs and GAMs to carrier loss experience, validate model assumptions and lift curves against holdout data, and deploy rate plans to the production pricing engine — replacing the historical cycle of manual R/Excel GLM builds that took weeks with an actuarially governed ML pipeline that takes days.

Build and validate insurance pricing models using Akur8 or hyperexponential (hx Renew): use the platform's transparent ML engine to fit GLMs and GAMs to carrier loss experience, validate model assumptions and lift curves against holdout data, and deploy rate plans to the production pricing engine — replacing the historical cycle of manual R/Excel GLM builds that took weeks with an actuarially governed ML pipeline that takes days.[12],[13]

Where your edge is

The platform accelerates model execution; the actuary's irreplaceable contribution is the judgment calls around variable selection, credibility weighting, territorial relativities, and regulatory defensibility. Develop fluency in explaining ML model outputs to regulators and underwriters — the "black box" risk is highest in Gradient Boosting models, and transparent GLM/GAM variants require the actuary to justify the model form in rate filings.

AI is sitting alongside you hereRun quarterly IBNR reserve analyses using Akur8 Arius (formerly Milliman Arius) or equivalent actuarial reserving software: set up loss development triangles, select development factors, apply chain-ladder and Bornhuetter-Ferguson methods, run stochastic reserve variability analyses, and produce the reserve exhibits required by GAAP, STAT, and IFRS 17 reporting — a workflow where AI platforms have reduced quarterly file-preparation time by up to 10× vs

Run quarterly IBNR reserve analyses using Akur8 Arius (formerly Milliman Arius) or equivalent actuarial reserving software: set up loss development triangles, select development factors, apply chain-ladder and Bornhuetter-Ferguson methods, run stochastic reserve variability analyses, and produce the reserve exhibits required by GAAP, STAT, and IFRS 17 reporting — a workflow where AI platforms have reduced quarterly file-preparation time by up to 10× vs. manual Excel-based reserving.[14],[15]

Where your edge is

AI-assisted reserving platforms automate the mechanical execution of triangle runs and variance analyses; the actuary's value is in the judgment-intensive steps: selecting the appropriate development method for each line (given data credibility and maturity), detecting anomalies in development patterns that signal claim operations changes or mix shifts, and translating reserve variability into business-language risk communication for CFOs and audit committees.

Where this role is heading

Natural next steps for someone with your foundation: not exits, evolutions.

A direction you could grow

Financial Managers

Senior actuaries — particularly Chief Actuaries and appointed actuaries at mid-size carriers — frequently transition into CFO or Financial Manager roles within insurance. The actuarial background provides direct preparation for insurance financial management: GAAP/STAT reserve adequacy, investment portfolio management under insurance liability constraints, reinsurance program oversight, and ORSA (Own Risk and Solvency Assessment) governance are all actuarial functions that scale naturally into financial leadership. Carriers actively look for actuarial-credentialed CFOs who can bridge quantitative risk modeling and financial reporting. The CRI is marginally higher because Financial Managers have a broader organizational mandate and the financial management credential moat (CPA, CFA, or equivalent) is complementary to the FSA/FCAS rather than redundant.

What you'd add
  • · GAAP and STAT insurance financial statement analysis: combined ratio management, investment income attribution, loss and LAE ratio decomposition
  • · Investment portfolio management under insurance liability constraints: duration matching, credit quality governance, NAIC investment category compliance
  • · ORSA and Solvency II / RBC capital framework: economic capital modeling, stress testing, regulatory capital reporting
  • · Executive communication: board-level risk appetite framing, investor relations for publicly traded carriers, rating agency engagement
  • · CPA or CFA credential as financial leadership credential complement to FSA/FCAS
What it takesSome new skills to pick up
Share this year
Drops anyone you send it to straight into 2026.
Preview card
Part of Tech & Data · see all 28roles →
Different role?

See the same long-arc view for your own profession.

Browse the directory by industry, or search by title or SOC code. New roles ship every few weeks. Every profile cites every claim.

Browse all roles

The data behind this timeline

On record since1762
Latest tracked employment26,670 (US, 2025)
Latest median pay$130,000 (2025)
Outlook+22% by 2034 (BLS Occupational Outlook 2024-34)
View all 27 cited data points
YearUS employmentMedian annual paySource
19491,069n/aESTIMATE
19695,500n/aESTIMATE
198918,000n/aESTIMATE
200014,000$72,270BLS-OEWS
200314,680$72,520BLS-OEWS
200416,350$76,340BLS-OEWS
200515,770$81,640BLS-OEWS
200616,620$82,800BLS-OEWS
200718,030$85,690BLS-OEWS
200818,220$84,810BLS-OEWS
200917,940$87,210BLS-OEWS
201019,700$87,650BLS-OEWS
201119,590$91,060BLS-OEWS
201221,340$93,680BLS-OEWS
201320,080$94,340BLS-OEWS
201425,648$96,700BLS-OEWS
201519,770$97,070BLS-OEWS
201619,940$100,610BLS-OEWS
201719,210$101,560BLS-OEWS
201820,760$102,880BLS-OEWS
201922,260$108,350BLS-OEWS
202022,480$111,030BLS-OEWS
202128,300$105,900BLS-OEWS
202225,010$113,990BLS-OEWS
202325,470$120,000BLS-OEWS
202433,600$125,770BLS-OEWS
202526,670$130,000BLS-OEWS
Embed this timeline on your site

Free for any site. Paste this where the timeline should appear; it stays interactive, every datapoint stays cited, and it sets no cookies on your page. How embedding works

<iframe src="https://futurehistory.earth/embed/15-2011"
  width="100%" height="430" style="border:0"
  title="Actuaries, a Future History timeline"
  loading="lazy"></iframe>

See all roles in Tech & Data