How to read an AI risk score

A number that claims to say how "exposed" your job is to AI can feel like a verdict. It isn't. It's one measurement, made a specific way, and it tells you something useful once you know how to read it. Start with the research, not the headlines.

What an exposure score actually measures

The most widely used measure comes from researchers Felten, Raj and Seamans, who linked progress in AI abilities — things like understanding language and recognising images — to the everyday abilities hundreds of different jobs rely on (read the peer-reviewed paper).

A higher score means more of your work overlaps with things AI is getting good at. That is all it means. It points to where AI will show up in your work — it does not say your job is disappearing. In many jobs, high overlap means AI takes on the repetitive slice and leaves the judgment, care, and craft to you.

Three questions to ask any AI claim about your job

• Can you click through to a real source? A claim you can check beats a confident claim every time. Genuine research and official data are public — you should be able to reach them in one click.
• Does it separate "AI helps with a task" from "AI takes over the work"? Those are very different futures, and honest analysis names which one it means.
• Is any growth or decline claim backed by official numbers? The U.S. Bureau of Labor Statistics publishes employment projections for hundreds of jobs. If a prediction doesn't trace back to numbers like these, treat it as opinion.

Where to check for yourself

• O*NET OnLine — the U.S. Department of Labor's public library of what each job really involves, kept current by surveying people who do the work.
• CareerOneStop — free career-planning tools from the Department of Labor, built on that same official data.
• Your free report here — we show how AI is affecting your specific role, with every claim linked to a real, working source.

Lead with what you already do well

A score can't see your strengths — the judgment, relationships, and hands-on skill you've built over years. The most useful next step is rarely to start over. It's usually to pair something you're already good at with one new AI-era skill, and let the number be background information instead of a headline.