JevMade hello@JevMade.com
← Back to guides

JevMade field notes / Data-pipeline walkthrough

A guide to scoring how AI changes thousands of job skills

You will learn how a project uses AI assistants to score thousands of job skills for automation risk. It shows how to break down careers into specific tasks to see where technology helps or takes over.

Original by Joris de VreedeEvaluation

Listen to this guide

JevMade’s plain-English explanation

0:00 /

AI narration

Credits

“AI-ISCO: Job Evolution Explorer” by Joris de Vreede. Read the original source.

This expanded guide is an AI-narrated adaptation prepared by JevMade. It expands the source’s essential ideas, examples and caveats in JevMade’s own words and is not a word-for-word reading. The synthetic voice does not imitate the author or imply their endorsement.

Our summary

This project explores how artificial intelligence changes over three thousand European careers. Instead of guessing how a whole job might change, it breaks each profession down into specific skills. Someone would use this method to see exactly which daily tasks technology might replace or improve.

The software feeds thousands of official job skills to an AI assistant, asking it to rate each one on a ten-point scale for automation risk and helpfulness. It also tests Jev, a different AI tool that chooses from set options rather than writing out text, to compare the scores.

These scores are estimates based on software tests, not measurements of real workplaces. The two AI tools disagreed on whether physical machinery counts as automation, showing that the scoring rules were unclear. This guide is useful for researchers building systems to evaluate large amounts of career data.

Key takeaways

  1. Break jobs down into individual skills before asking an AI assistant to score them for automation.
  2. Write clear rules for your one-to-ten scales so the software understands exactly what each number means.
  3. Save your progress frequently when processing thousands of items so you do not lose work during errors.

The guide names specific AI models used for the scoring tests, but the reviewer did not rerun the results. The scores represent a reproducible software process rather than independently confirmed facts about the job market.

Repository README · Source reviewed

Read the original guide Opens the author’s site in a new tab.