Will AI replace Data Scientists?

AI can do 95% of the work Data Scientists do at least twice as fast. People already use it for 1 of their 16 tasks, while 14 more are ready but barely used so far. 1 still needs a person.

AI is already doing this1 · 7% of time

People already use AI for these tasks in real work, based on Anthropic's analysis of how Claude is used.

  • Identify relationships and trends or any factors that could affect the results of research.Mostly assisted7% of time

AI could do this next14 · 88% of time

AI can speed these up by at least half, but real use has not caught up yet. These are next in line.

  • Analyze, manipulate, or process large sets of data using statistical software.15% of time
  • Clean and manipulate raw data using statistical software.15% of time
  • Write new functions or applications in programming languages to conduct analyses.11% of time
  • Test, validate, and reformulate models to ensure accurate prediction of outcomes of interest.9% of time
  • Create graphs, charts, or other visualizations to convey the results of data analysis using specialized software.7% of time
  • Deliver oral or written presentations of the results of mathematical modeling and data analysis to management or other end users.6% of time
  • Identify business problems or management objectives that can be addressed through data analysis.5% of time
  • Identify solutions to business problems, such as budgeting, staffing, and marketing decisions, using the results of data analysis.5% of time
  • Recommend data-driven solutions to key stakeholders.5% of time
  • Apply feature selection algorithms to models predicting outcomes of interest, such as sales, attrition, and healthcare use.4% of time
  • Read scientific articles, conference papers, or other sources of research to identify emerging analytic trends and technologies.4% of time
  • Apply sampling techniques to determine groups to be surveyed or use complete enumeration methods.1% of time
  • Design surveys, opinion polls, or other instruments to collect data.1% of time
  • Propose solutions in engineering, the sciences, and other fields using mathematical theories and techniques.1% of time

Still needs a person1 · 5% of time

No real AI use, no proven AI speedup and no robot that can do it yet. This is the part of the job to build on.

  • Compare models using statistical performance metrics, such as loss functions or proportion of explained variance.5% of time

Your week is not the average

Describe what you actually do and get a personal report: your own tasks, what AI can already do, what stays yours and a 90 day plan.

Get my personal report

What this means

AI can already do almost all of this work at least twice as fast. Real use is still behind at 5 out of 100, so the change here is only getting started. Expect the role to shift toward deciding what to do, checking AI's work and owning the result.

What to do next

  • Do the 1 task AI is already handling with AI yourself. Your peers already are, and it is quickly becoming the baseline.
  • Get ahead on the 14 tasks AI could do next. These move as tools improve, so learning them early pays off.
  • Build on the 1 task that stay human. This is where your judgment, skill and relationships matter most.
Business Intelligence Analysts95%Clinical Data Managers84%Biostatisticians93%Operations Research Analysts98%Actuaries99%Statisticians99%

Get the weekly brief

One short email a week on how AI is changing work, including jobs like this one.

Check another job