- People at highest risk from AI may be the least prepared to adapt... who are they? What can be done? *
“The people most at risk of being replaced by AI are often the least resourced to adapt. What if we could identify them — and help?”
This project explores which occupations and communities are most vulnerable to AI-driven displacement using public datasets from O*NET, the Bureau of Labor Statistics, and the U.S. Census. The goal is to uncover patterns of risk, highlight skill retraining pathways, and propose data-driven interventions that could improve the lives of millions.
- Score U.S. jobs by AI susceptibility
- Identify at-risk populations using geographic and demographic overlays
- Explore skill transferability and upskilling opportunities
- Publish findings to spark conversation and collaboration
- Python, pandas, SQL (via dbt)
- Causal inference & clustering
- Geospatial visualization
- Streamlit for interactive exploration
- GitHub for transparency & reproducibility
- Vector embedding for U.S. jobs to understand semantic distance between different classes of AI susceptible jobs
data/: Raw and processed datasetsnotebooks/: Analysis and modelingstreamlit_app/: Optional frontend for interactivityreports/: Whitepaper, data visuals, final artifacts
Millions face the threat of job loss due to automation, yet very few have a path forward. If we can spotlight risk with clarity and compassion — and show what's possible with data — we can spark policy, products, and partnerships that actually help.
- Vulnerability Index by Occupation
- Map of AI Disruption Hotspots
- Retraining Opportunity Graph
- Public Report + Streamlit Dashboard
- Video Explainer: “What if we prepared instead of panicked?”
This repo is a work in progress. If you're an engineer, designer, educator, policymaker, or just curious human — let’s jam. DM me or open an issue.