Andrej Karpathy released an interactive AI Job Exposure Map analysing 342 occupations and roughly 143 million US jobs to estimate how vulnerable different roles are to automation.
The experiment used job descriptions from the US Bureau of Labor Statistics and fed them into a large language model that assigned exposure scores from zero to 10 based on how much AI could potentially reshape each role.
Across the workforce the weighted average exposure score was about 4.9, while roughly 42% of jobs — about 59.9 million workers earning an estimated $3.7 trillion in annual wages — scored seven or higher.
High-paying knowledge roles including lawyers, accountants, financial analysts and software developers ranked among the most exposed occupations because their work involves structured digital information and research tasks.
In contrast, hands-on professions such as plumbers, electricians and construction labourers recorded some of the lowest exposure scores, reflecting the difficulty of automating unpredictable physical work.
Reacting to the viral chart, Elon Musk commented on X that:
“All jobs will be optional. There will be universal high income.”
Karpathy later removed the website and GitHub repository for the project, describing the visualisation as a quick experimental exercise, though copies quickly spread online as developers archived and replicated the dataset and tools.