According to Analytics India Magazine, Andrej Karpathy mapped the AI exposure of jobs across 342 occupations in a viral weekend project published on March 16, 2026. The OpenAI co-founder analysed 143 million US jobs using Bureau of Labor Statistics data to estimate how artificial intelligence could influence workplace tasks.
Andrej Karpathy AI job analysis examines 342 occupations
The AI exposure of jobs became widely discussed after Andrej Karpathy released an interactive visualisation on March 15. The tool evaluated 342 occupations listed by the Bureau of Labor Statistics.
Karpathy created the project as a short experiment. He said he completed the analysis in roughly two hours during a Saturday morning session.
Soon after publication, the visualisation gained traction online. However, Karpathy later removed the website after some readers interpreted the results as predictions of job losses.
AI exposure of jobs highlights digital vs manual roles
The AI exposure of jobs analysis used a scoring scale from 0 to 10. A score of 10 indicated the highest potential influence from artificial intelligence tools.
The study calculated a weighted average score of about 4.9 across the US labour market. Approximately 42 percent of jobs nearly 59.9 million roles generating around $3.7 trillion in wages scored seven or higher.
Higher-income professions recorded stronger exposure. Jobs earning more than $100,000 per year averaged 6.7 on the scale. In contrast, occupations earning below $35,000 averaged around 3.4.
AI exposure of jobs shows differences across occupations
Digital-focused professions ranked among the highest in the AI exposure of jobs map. Software developers, data analysts, accountants, lawyers, financial specialists, writers and medical transcriptionists scored between eight and ten.
Manual roles showed lower exposure levels. Occupations such as roofers, janitors and home health aides typically ranked near one or below on the scale.
Karpathy clarified that the exposure scores reflected how digital a job’s tasks are. He noted that the estimates did not predict employment losses and did not include factors such as demand elasticity, regulation or social preferences.
AI exposure of jobs sparks wider discussion
The online discussion around the AI jobs analysis expanded after technology leaders reacted publicly. Tesla and SpaceX CEO Elon Musk commented that advances in artificial intelligence could eventually make work optional.
Meanwhile, Karpathy stated that the tool was designed to help researchers explore labour data visually. He encouraged others to modify the dataset or develop new visualisations based on the same information.
