Will AI replace data scientists?
Updated
Data scientists score 67 of 100 for AI exposure. The largest share of their task time, 62%, is work AI speeds up while they stay in charge. AI can take over the other 38% outright.
- 38% AI can do it
- 62% AI helps
High exposure
Higher than 98% of the 923 occupations we scored.
Task by task
We graded 16 tasks, listed here from most to least exposed. AI can do 6 of them today, speeds up 10 and leaves 0 to people.
| Task | AI exposure |
|---|---|
| Clean and manipulate raw data using statistical software. Cleaning and reshaping raw data is highly automatable with code assistants. | 80% AI can do it |
| Write new functions or applications in programming languages to conduct analyses. Code assistants write most analysis functions; developer reviews and tests output. | 80% AI can do it |
| Create graphs, charts, or other visualizations to convey the results of data analysis using specialized software. AI generates charts and dashboards code quickly; human checks accuracy. | 75% AI can do it |
| Analyze, manipulate, or process large sets of data using statistical software. AI coding assistants can write and run most data processing scripts. | 72% AI can do it |
| Read scientific articles, conference papers, or other sources of research to identify emerging analytic trends and technologies. LLMs summarize papers and surface trends quickly; reviewer checks relevance. | 65% AI can do it |
| Compare models using statistical performance metrics, such as loss functions or proportion of explained variance. Metric computation is routine; AI can compare and explain tradeoffs. | 62% AI can do it |
| Apply feature selection algorithms to models predicting outcomes of interest, such as sales, attrition, and healthcare use. Libraries and AI suggest feature selection; analyst judges business relevance. | 58% AI helps |
| Identify relationships and trends or any factors that could affect the results of research. AI finds correlations quickly; judging causal factors needs domain expertise. | 58% AI helps |
| Test, validate, and reformulate models to ensure accurate prediction of outcomes of interest. AI runs validation checks and suggests fixes; analyst is responsible for acceptance. | 55% AI helps |
| Design surveys, opinion polls, or other instruments to collect data. AI drafts questions and reviews wording; survey design needs research judgement. | 50% AI helps |
Show the other 6 tasks
| Task | AI exposure |
|---|---|
| Apply sampling techniques to determine groups to be surveyed or use complete enumeration methods. AI can compute sample sizes and draw samples; design needs judgement. | 45% AI helps |
| Propose solutions in engineering, the sciences, and other fields using mathematical theories and techniques. AI helps derive and check models; field validation remains expert led. | 40% AI helps |
| Deliver oral or written presentations of the results of mathematical modeling and data analysis to management or other end users. AI drafts slides and summaries; delivering and fielding questions stays human. | 35% AI helps |
| Identify solutions to business problems, such as budgeting, staffing, and marketing decisions, using the results of data analysis. AI can model options; choices involve budgets, people and political tradeoffs. | 35% AI helps |
| Recommend data-driven solutions to key stakeholders. AI drafts recommendations; stakeholder trust and accountability stay with analyst. | 30% AI helps |
| Identify business problems or management objectives that can be addressed through data analysis. Problem framing depends on stakeholder conversations and organizational context. | 22% AI helps |
How data scientists compare
All occupations (923)
Computer and mathematical (36)
Pay and outlook in the US
- Median pay
- $120,230 a year, May 2025
- People employed
- 262,440 in the US, May 2025
- Projected change in jobs
- +34.6% from 2024 to 2034
Skills to build
- Frame business problems with stakeholders before any coding
- Review AI-written Spark and Hive jobs against source data
- Orchestrate scheduled pipelines in Apache Airflow
- Prepare and blend data workflows in Alteryx software
- Present model tradeoffs to management with written decision memos
- Design sampling plans for surveys and experiments
Safer roles nearby
5 related occupations with lower scores.
Questions about data scientists and AI
Will data science be replaced by AI?
Partly. About 5 in 10 working hours go to tasks AI can already do, mostly cleaning raw data and writing analysis code. Problem framing, stakeholder recommendations and presenting results stay with the person who answers for the decision.
Will AI take over data science?
AI can already do 38% of the task time in data science and speeds up another 62%. The US has 262,440 data scientist jobs, with a projected change of +34.6% from 2024 to 2034.
Which tasks is AI taking over from data scientists?
Cleaning and reshaping raw data goes first, because code assistants can write the transformation scripts from a short description. Writing analysis functions and building charts follow, with exposure scores of 80 and 75 out of 100. The analyst still reviews and tests each output before it reaches a decision.
Which tasks stay with data scientists?
Framing the business problem stays with the analyst, since it depends on stakeholder conversations and organizational context. Recommending solutions, presenting results and choosing between budget, staffing and marketing options also stay human. Trust and accountability sit with the person who answers for the decision.
How can data scientists stay valuable as AI takes over?
Move hours toward problem framing and stakeholder sessions, the least exposed tasks in this job. Review every AI-written Spark or Airflow job against sample source data before it runs. Take responsibility for the written decision memo that goes to management with each model result.
Related jobs
Next steps for data scientists
Check which roles in your team AI can take on
We map your team's tasks, show what AI can do today and build the automations that pay back first.
Related topics
- Jobs AI can't replaceCement masons score lowest, at 8.
- Jobs most at risk from AIProofreaders score highest, at 76.
- High-paying jobs with low AI exposureRadiologists earn a median $420,860 and score 46.
- Entry-level jobs with low AI exposureCement masons need little prior training and score 8.
- Jobs AI speeds up the mostFor natural sciences managers, AI speeds up 96% of task time.
- AI and jobs in Malta15 iGaming roles graded task by task.
- Skills that keep work humanThe skills to build in the jobs AI changes least.
- Jobs A to ZEvery job we publish, by first letter.
- AI exposure by industryAverage scores for 22 occupation groups.
- Write a CV screening software can readHow applicant tracking systems read a CV, and what to change.
- AI for businessesThe tasks to automate first in each team and industry.
- How we score AI exposureThe grading, the weights and every source.