← Explore professions

Technology

Data Scientist

Based on 40 assessments · 4 from real users

34% Moderate risk

Average realistic automation risk across all Data Scientist profiles in the dataset.

Raw potential
78%
Realistic risk
34%
Research benchmark ?
59%

Raw potential = I/O automation ceiling. Realistic risk = adjusted for informal knowledge and social context. Research benchmark: Eloundou et al. (2023)

Distribution across 40 profiles. Middle half of Data Scientists score between 30% and 35%.

0% 50% 100%
p10 · 28%
37% · p90
On-screen work 76%

Done entirely on a computer. High AI exposure — these tasks are already in the automation zone.

In-person + screen 13%

Physical sensing, digital output — e.g. interviewing someone then writing a report. Partially protected.

Computer + action 10%

Computer input, real-world output — needs someone to act on it, not just software.

Fully in-person 1%

No computer required. Furthest from automation — the strongest human advantage.

3 synthetic profiles for a Data Scientist, ordered by automation exposure. Tab between them to see how task mix drives the score difference.

Task Time Type Exposure
Developing and training machine learning models (e.g., regression, classification, clustering) to solve business problems or make predictions
deep expertise social element
25% DD 29%
Exploratory data analysis (EDA) to identify patterns, trends, or anomalies in datasets using statistical methods and visualizations
deep expertise social element
21% DD 21%
Cleaning and preprocessing raw data (e.g., handling missing values, removing duplicates, standardizing formats) to prepare it for analysis
deep expertise
18% DD 37%
Collaborating with stakeholders (e.g., product managers, engineers) to define project goals, interpret results, and align models with business needs
deep expertise
16% AD 6%
Creating reports, dashboards, or presentations to communicate insights, model performance, or recommendations to non-technical audiences
deep expertise social element
9% DA 6%
Optimizing model performance through hyperparameter tuning, feature engineering, or selecting alternative algorithms
deep expertise
7% DD 38%
Monitoring deployed models in production to track performance, detect drift, and retrain models as needed
1% DD 62%

Work as a Data Scientist? Map your specific role.

Start assessment →