Executive Summary

In Magnifica Humanitas, Pope Leo XIV urged the world to safeguard the human person in the time of artificial intelligence. He grounded human dignity in our innate capacity for relationship and communion, warning that when people are treated as objects to be perfected or surpassed, it becomes easier to regard some lives as less useful, less desirable, or less worthy.

The Magnificent Humanity Labor & Flourishing Index is the Foundation's answer to that call in the one arena where AI's impact on human dignity is most immediate and measurable: work. It takes the Holy Father's charge seriously by refusing to treat labor as a mere economic input — and building an instrument rigorous enough to prove, not just proclaim, whether people are flourishing.

The Magnificent Humanity Foundation’s Labor & Flourishing Index is a five-step intervention framework for helping families whose livelihoods have been disrupted by artificial intelligence.

It works by:

1. Tracking AI-related job displacement in real time.

2. Identifying the affected families behind the employment data.

3. Vetting applicants to confirm that their displacement genuinely resulted from AI.

4. Stabilizing eligible families through a 90-day “Human Dignity Dividend,” providing temporary financial security.

5. Re-educating recipients with skills intended to lead toward more resilient employment.

Its purpose is to detect displacement early and move families from financial shock toward stability, retraining, and a sustainable livelihood. In short, it aims to turn labor-market data into direct, human-centered assistance.

The problem nobody is answering. Every serious labor-market study measures employment, wages, productivity, or automation risk. None of them ask the question that actually matters: are people, families, and communities better off because of AI? Are they flourishing? That's the gap. Employment statistics tell you what happened to a job. They don't tell you what happened to a person. This Index is built to answer that second question first.

What we can build right now, with public data alone. By pulling O*NET task data, BLS wage and employment figures, Census/ACS demographics, and IPEDS training-pipeline data, we can produce a scorecard — four pillars, scored 0 to 100, legible to a policymaker or donor in five minutes:

• AI Exposure — how much of this work can AI automate or augment

• Worker Vulnerability — how many workers are affected, and how economically fragile they are

• Transition Capacity — whether workers can realistically move into better roles

• Flourishing Opportunity — where targeted support does the most good

This is a map of risk and opportunity, not a prediction. Nothing manufactured, nothing overstated — a defensible baseline built entirely on public data, refined using AI to classify and score at a scale no manual team could match.

What only the Foundation can build — and where donor capital becomes decisive. Public data can describe exposure. It cannot prove what actually happened to a specific worker, whether retraining worked, or whether someone is flourishing on the other side of displacement. That proof requires original research: a Worker Transition Survey, an Employer AI Adoption Survey, a Human Flourishing Survey, and a longitudinal dataset tracking the Foundation's own scholarship and grant recipients. This is proprietary evidence that exists nowhere else, and it's what philanthropic capital is for.

The path. Four phases over 24 months — public-data prototype, AI-refined scorecard, first proprietary surveys with partners, and an Annual Index tracking Foundation beneficiaries over time. Each phase produces a real deliverable, not a status update.

Why it will hold up. The Index runs on hard guardrails: exposure is never confused with displacement, every estimate carries a confidence level instead of false precision, and no claim is made without evidence behind it. That discipline is what makes the Index usable — by policymakers, by employers, and by the Foundation itself, as an allocation tool that directs grants, scholarships, and partnerships toward what demonstrably helps, not what merely sounds good.

The result, at every stage: does AI leave humanity better off. Everything else is in service of answering that.

By pairing free public datasets with AI-driven analysis, we can show precisely where workers, occupations, industries, regions, and institutions face the greatest risk from AI disruption, and where the greatest opportunity for human flourishing lies. No new funding is required to produce the first version. It requires only the discipline to build it well.

We begin with a transparent, evidence-based map of risk and opportunity. Nothing manufactured, nothing overstated. As the Foundation grows, that map becomes a living instrument, sharpened continually by proprietary survey data drawn directly from workers, employers, educators, and the people our programs serve. This is where donor support becomes decisive: public data can describe exposure, but only original research can prove what actually happens to people, and what actually helps.

Human Flourishing: The Guiding Purpose

Every serious labor-market study measures employment, wages, productivity, or automation risk. This Index asks a harder question, one no one else is asking with the same rigor: are people, families, and communities actually better off because of AI?

This is the differentiator. Inspired by Pope Leo XIV's Magnifica Humanitas and the vision of the Magnificent Humanity Pledge, the Index treats labor not as a mere economic input but as an expression of human creativity, contribution, purpose, and dignity. Every person carries inherent worth that no algorithm assigns and no automation can revoke.

Others already publish employment statistics, automation forecasts, and workforce studies. Our contribution is to fuse those traditional labor indicators with direct measures of human flourishing — economic mobility, educational opportunity, purpose, optimism, and community resilience — so we can answer not only "What is happening to jobs?" but "What is happening to people?" That second question is the one policymakers, employers, and funders cannot currently answer. It is the one this Index is built to answer first.

In every phase of development, the Index returns to one governing question: does artificial intelligence leave humanity better off? The labor market offers the earliest, clearest signal — but it is only the opening chapter. The Foundation's long-term purpose is to identify where AI creates risk, where it creates opportunity, and which investments, policies, and philanthropic initiatives most effectively advance human flourishing.

The Big Idea, in Plain English

What we can already know from public data: which jobs are exposed to AI, how many people work in them, where they live, what they earn, and what retraining capacity exists nearby — mapped across occupational risk, regional vulnerability, education gaps, wage trends, and transition pathways.

What only the Foundation can create: proof. Whether a specific worker lost a job because of AI, whether retraining actually worked, and whether people are flourishing afterward — captured through worker transition stories, employer AI-adoption detail, real displacement attribution, and direct flourishing outcomes. This is proprietary evidence no public dataset contains, and it is what philanthropic capital makes possible.

Step-by-Step Methodology

Step 1 — Define the mission

State plainly that the Index measures both risk and opportunity: who may be harmed by AI disruption, who may benefit from AI augmentation, and where support can help people flourish.

Step 2 — Start with occupations

Occupations are the foundation because they are data-rich and connect cleanly to wages, employment, regions, industries, and education pathways. Every occupation score can later roll up into industries, regions, companies, and grant-making priorities.

Steps 3–5 — Score every occupation's AI exposure

Step 3: Pull O*NET data on job tasks, skills, abilities, knowledge, and education requirements — the authoritative public record of what people actually do in each occupation.

Step 4: Use AI to evaluate each task and classify it as automatable, augmentable, or protected. Routine paperwork may be automatable; complex client advising, augmentable; hands-on caregiving, more protected.

Step 5: Combine task-level scores into a single occupation-level score — the first version of the Index. Not proof of job loss. A reasonable, defensible map of exposure.

Method note: This phase relies on descriptive statistics, composite index construction, and standardization — deliberately simple methods, because at this stage the Foundation is measuring and describing exposure, not predicting job losses. A reliable baseline first; predictive claims come later, once they can be earned.

Steps 6–9 — Add wages, geography, and readiness

Step 6: Layer in BLS employment and wage data — how many workers hold each job, what they earn, and whether the occupation is growing or shrinking.

Step 7: Layer in Census/ACS geography to find where exposed workers live and which communities carry higher vulnerability due to income, education, age, commuting patterns, or limited retraining access.

Step 8: Layer in NCES/IPEDS data to map nearby colleges, certificate programs, and training pipelines that could help workers transition.

Step 9: Publish the first public Index — a scorecard clear enough for a policymaker, an employer, or a donor to read in five minutes and act on.

Method note: This phase adds geographic analysis, cluster analysis, and, where useful, principal component and factor analysis — methods built to reveal real patterns of vulnerability and opportunity, not to manufacture false certainty.

Step 10 — Let AI do the heavy lifting

AI can process thousands of occupational tasks, summarize trends, read public filings, detect AI-related workforce announcements, cluster similar occupations, and suggest retraining pathways — at a speed and scale no manual team could match.

Method note: Predictive analytics, regression models, time-series forecasting, and machine learning classification help forecast disruption and flag opportunities before they fully emerge — turning the Index from a snapshot into an early-warning system.

Step 11 — Move from public estimate to proprietary evidence

Once the public-data version exists, launch surveys to answer what no government dataset can: actual worker displacement, employer adoption decisions, retraining results, and flourishing outcomes. This is the point at which the Index stops describing the labor market in general and starts documenting what is happening to real people — and it is where donor investment converts directly into original knowledge the world does not yet have.

Method note: Inferential statistics, confidence intervals, hypothesis testing, and proper survey sampling let the Foundation move from estimates to statistically valid conclusions about displacement, adoption, and outcomes — evidence that can stand up to scrutiny from journalists, policymakers, and skeptics alike.

Step 12 — Turn the Index into action

Use the results to direct grants, scholarships, employer partnerships, policy recommendations, education programs, and investment-fund priorities — so the Index is never just a report. It is an allocation tool for capital that wants to do the most good.

Method note: Difference-in-differences, panel models, survival analysis, and treatment-effect estimation determine whether AI actually caused displacement and whether an intervention worked — the difference between funding what feels good and funding what demonstrably helps.

The Version 1 Scorecard: Four Pillars

Every occupation and region receives a plain, 0-to-100 score across four pillars — legible to a policymaker, an employer, or a first-time donor alike:

• AI Exposure — How much of this work could AI automate or augment? Built from O*NET task and skill data, with AI classifying exposure.

• Worker Vulnerability — How many workers are affected, and how economically fragile are they? Built from BLS wage and employment data plus ACS demographics, ranked for risk.

• Transition Capacity — Can workers realistically move into better roles? Built from O*NET skill adjacency and IPEDS training capacity, mapping real retraining paths.

• Flourishing Opportunity — Where can targeted support do the most human good? Built from income, education, region, and local training assets, to recommend grant and scholarship priorities.

What Must Become Proprietary — Where Your Support Matters

Public data can only take the Index so far. The following instruments do not exist anywhere else, and building them is the Foundation's core research investment:

• Worker Transition Survey — captures whether people were displaced, whether AI was a factor, how long they were unemployed, whether they retrained, and whether income recovered.

• Employer AI Adoption Survey — captures whether employers used AI to cut headcount, freeze hiring, redesign roles, retrain workers, or create new positions.

• Human Flourishing Survey — measures income stability, purpose, optimism, community connection, family stability, and educational opportunity after disruption.

• Foundation Program Outcomes Dataset — tracks scholarship and grant recipients over time to learn, with evidence, what interventions actually help people flourish.

Building the Index: A Four-Phase Timeline

• Phase 1 (Months 0–2): Build the public-data prototype from O*NET, BLS, Census/ACS, and IPEDS. Deliverable: the first exposure and vulnerability scorecard.

• Phase 2 (Months 2–3): Use AI to refine task scoring, transition pathways, and regional rankings. Deliverable: a stakeholder-ready report and public methodology.

• Phase 3 (Months 3–6): Launch small worker and employer surveys with partners. Deliverable: the first proprietary displacement and transition data released to the public.

• Phase 4 (Months 6+): Expand surveys and track Foundation beneficiaries longitudinally. Deliverable: the Annual AI Labor & Flourishing Index.

Guardrails That Protect Our Credibility

A donor's trust is the Foundation's most valuable asset, and it is protected by discipline, not marketing. The Index will hold to these rules without exception:

• Never claim exact AI displacement counts until the Foundation has direct evidence or a transparent attribution model.

• Separate exposure from displacement. Exposure means a job can be affected by AI; displacement means jobs, hours, or wages were actually lost because of it.

• Publish confidence levels. A plain high/medium/low label beats false precision every time.

• Keep the Index constructive, never punitive. The goal is human flourishing — better retraining, responsible adoption, targeted support.

• Treat privacy as non-negotiable. Worker surveys collect only what is necessary and are aggregated before publication.

Appendix A: Public Data Sources

The first version of the Index is built entirely on free, authoritative public data:

• O*NET Resource Center / Database — occupational tasks, skills, abilities, knowledge, and education requirements. onetcenter.org/database.html

• O*NET Web Services — API access to the O*NET database for automated workflows. services.onetcenter.org

• U.S. Bureau of Labor Statistics Public Data API — published historical time series in JSON or Excel. bls.gov/bls/api_features.htm

• BLS Occupational Employment and Wage Statistics — annual employment and wage estimates for roughly 830 occupations, nationally and by state, metro, and industry. bls.gov/oes

• U.S. Census Bureau ACS API — demographics, income, education, and geography. census.gov/programs-surveys/acs/data/data-via-api.html

• IPEDS / NCES Use the Data — data on more than 7,000 institutions, downloadable as custom or complete files. nces.ed.gov/ipeds/use-the-data

Appendix B: Statistical Methods in Plain English

For donors and board members who want to understand the rigor behind each number, here is what each method means, in plain terms — with one example from the Index and one from everyday life.

Descriptive Statistics

Summarizing and describing data, to understand what is happening right now. Index example: the average wage and number of workers in highly AI-exposed occupations. Everyday example: the median home price in California.

Composite Score Construction

Combining several measurements into one clear score. Index example: blending AI exposure, wage vulnerability, and retraining difficulty into a single Worker Vulnerability Score. Everyday example: a credit score combining payment history, debt, and account age.

Standardization and Normalization

Putting different measurements on the same scale so they can be compared. Index example: converting wages, unemployment, and retraining capacity into comparable 0–100 scores. Everyday example: converting SAT and ACT scores onto a common scale.

Inter-Rater Reliability Testing

Checking whether independent reviewers reach the same conclusion. Index example: comparing AI models and human experts scoring whether an occupation is automatable. Everyday example: three doctors independently reading the same X-ray.

Trend Analysis

Identifying whether a pattern is rising, falling, or holding steady. Index example: tracking whether employment in highly exposed occupations declines year over year. Everyday example: tracking annual growth in electric-vehicle sales.

Rolling Averages

Smoothing short-term noise to see the real trend. Index example: a 12-month moving average of layoffs in AI-exposed occupations. Everyday example: the stock market's 200-day moving average.

Geographic and Cross-Sectional Comparisons

Comparing groups or places to find who is most vulnerable or resilient. Index example: comparing AI exposure in California versus Kentucky. Everyday example: comparing diabetes rates across states.

Cluster Analysis

Automatically grouping similar cases to reveal hidden patterns. Index example: grouping occupations with similar AI risks and retraining paths. Everyday example: Netflix grouping viewers with similar tastes.

Predictive Analytics

Using data to forecast what happens next. Index example: predicting which occupations are likeliest to see AI-driven job losses. Everyday example: forecasting retail demand before the holidays.

Regression Models

Estimating how one factor influences another. Index example: estimating how AI exposure affects wages or employment. Everyday example: estimating how education affects income.

Machine Learning Classification

Sorting cases into categories automatically, to sharpen prediction. Index example: classifying occupations as low, medium, or high risk for AI disruption. Everyday example: credit-card fraud detection.

Bayesian Updating

Revising an estimate as new evidence arrives. Index example: raising the probability of displacement once new employer surveys show widespread AI adoption. Everyday example: updating a weather forecast as new satellite data comes in.

Inferential Statistics

Using a sample to draw conclusions about the whole population, because surveying everyone isn't possible. Index example: surveying 5,000 workers to estimate national AI-displacement trends. Everyday example: political polling.

Longitudinal Analysis

Following the same people over time to measure lasting outcomes. Index example: tracking displaced workers for five years after retraining. Everyday example: the Framingham Heart Study.

Panel Models

Analyzing many groups across many time periods to separate temporary blips from real trends. Index example: following every occupation annually from 2025 to 2040. Everyday example: studying every U.S. state over several decades.

Survival Analysis

Studying how long it takes for an event to occur, because timing matters. Index example: measuring how long displaced workers remain unemployed before finding new work. Everyday example: studying time to cancer recurrence.

Treatment-Effect Estimation

Measuring whether an intervention actually caused an improvement. Index example: determining whether Foundation scholarships improve employment outcomes after displacement. Everyday example: evaluating whether a new medicine improves patient survival.

Sensitivity Analysis

Testing whether conclusions hold up when assumptions change, to avoid false precision. Index example: changing the Index's weights to see if occupational rankings shift. Everyday example: stress-testing mortgage assumptions.

Confidence Intervals

Showing the range of uncertainty around an estimate, because every estimate has some. Index example: estimating that 100,000–140,000 workers may be at risk, rather than claiming exactly 120,000. Everyday example: the margin of error in election polling.

Back-Testing and Out-of-Sample Validation

Testing a model on data it has never seen, to confirm it actually works. Index example: building a displacement model on older data, then testing it on recent labor-market data. Everyday example: testing an investment strategy against historical markets.

Difference-in-Differences

Comparing two groups before and after an event to estimate what the event actually caused. Index example: comparing highly exposed occupations to low-exposure ones before and after widespread generative-AI adoption. Everyday example: studying the effects of a minimum-wage change.

Prescriptive Analytics

Turning predictions into recommended action, because prediction alone isn't enough. Index example: recommending where the Foundation should deploy scholarships and grants. Everyday example: Google Maps recommending the fastest route.

Scenario Simulation

Modeling plausible "what-if" futures, because the future is uncertain. Index example: estimating what happens if AI adoption doubles over the next five years. Everyday example: insurers modeling hurricane scenarios.

The Magnificent Humanity Labor & Flourishing Index is the Foundation's answer to His Holiness Pope Leo XIV’s call in the one arena where AI's impact on human dignity is most immediate and measurable: work. It takes the Holy Father's charge seriously by refusing to treat labor as a mere economic input — and building an instrument rigorous enough to prove, not just proclaim, whether people are truly flourishing.