AI and Data scientists

—What it means

This sector is among the most exposed: much of its typical work overlaps with what AI already does. In practice: AI won't erase it, it'll reorganise it. Those who use it as a copilot and move up toward judgment, relationships and decisions — what AI can't close — will win.

AI already writes pipelines, tunes models and explores data automatically; what it doesn't take on is framing the right question, choosing the metric or judging whether the model actually serves the business.

What to do

• 01

  Automate the routine

  Hand the repeatable and predictable to AI to free time and attention for what only you can do.

• 02

  Augment your judgment

  Use AI as a copilot to do what you already do better — analyse, draft, explore ideas — with your judgment in charge.

• 03

  Anticipate what's next

  Grow what AI can't close: judgment, relationships, creativity and continuous learning. That's your edge.

FAQ

Will AI replace Data scientists?

Exposure measures task overlap, not replacement. Usually it's not disappearing but a change in what you do and how — toward judgment, relationships and what AI can't close. Exposure isn't destiny.

What does AI already do in Data scientists — and what doesn't it?

AI already writes pipelines, tunes models and explores data automatically; what it doesn't take on is framing the right question, choosing the metric or judging whether the model actually serves the business.

Where does this number come from?

Occupation-level estimate (not your exact tasks), based on AIOE (Felten et al., 2021) and “GPTs are GPTs” (Eloundou et al., 2024).