Frequently Asked Questions
Everything you need to know about how the index works, where the data comes from, and what the score means.
The Index
What is the Takeover Tracker?
It’s a daily composite index (0–100%) that estimates how far AI has progressed toward making white-collar work economically obsolete. Think of it like the Doomsday Clock, but for jobs — a single transparent number updated every day using public data and AI-powered analysis.
What does the score actually mean?
0% represents the pre-AI baseline (circa 2019). 100% would mean human white-collar labor offers no cost advantage over AI. The current estimate sits around 20%, reflecting established AI assistance and the beginning of entry-level task automation.
How often does the score update?
The pipeline runs daily at 6:00 AM UTC. It collects fresh data, analyzes it with AI, scores 24 sector–category pairs, and produces one number. The entire process takes about 15 minutes.
Can the score go down?
Yes. If labor market data improves, corporate AI adoption slows, or economic indicators shift, the score can decrease. A 2-point daily movement cap and dual-EMA smoothing prevent wild swings in either direction.
Methodology
Where does the data come from?
Nine categories of public sources: BLS and FRED labor/economic statistics, RSS feeds from major outlets, GDELT global events, Reddit and Hacker News discussions, SEC filings, AI benchmark results, Google Trends, and layoff trackers. Every source is documented on the Methodology page.
How is the score calculated?
Signals are scored across 4 job sectors and 6 signal categories (24 sub-scores). These are aggregated using a weighted geometric mean (like the UN’s HDI), adjusted for hype, and smoothed with a dual-EMA filter. The full formula is published in our methodology.
What stops the score from being inflated by AI hype?
A built-in hype resistance mechanism compares reality signals (labor data, corporate adoption, economic indicators) against hype signals (AI benchmarks, sentiment). When hype exceeds 2× reality, capability and sentiment scores are automatically discounted.
Is the methodology open?
Completely. Every data source, weight, formula, and AI prompt is published. The scoring rubric is editable and versioned. Anyone can audit, critique, or reproduce the results.
Features
Can I look up specific occupations?
Yes. We have displacement risk scores for 1,100+ occupations based on O*NET task and attribute data. You can search by job title and see task-level breakdowns of AI exposure.
Can I track industries?
Yes. Browse industries, view risk scores, and track the ones relevant to you from your account dashboard.
Is there a risk quiz?
Yes. The AI Impact Quiz asks about your role, tasks, and work environment, then estimates your personal displacement risk based on the same methodology that powers the index.
Is Takeover Tracker free?
The public dashboard, historical data, occupation lookups, and methodology are all free. We plan to offer premium features like alerts, API access, and detailed reports in the future.
Technical
What AI model powers the analysis?
Google’s Gemini (configurable, currently Gemini 2.0 Flash). Each scoring call runs 3 times at low temperature and takes the median, a technique called self-consistency sampling that filters out outlier responses.
Is there an API?
A public API is in development. Check the API Documentation page for current endpoints and usage details.
Can I contribute or give feedback?
Absolutely. The project is open to feedback. Visit our Contact page to get in touch, or check the changelog for recent updates.
Want the full picture?
The Methodology page documents every weight, formula, data source, and prompt used to produce the score.