AI automation risk varies significantly between digital and physical professions

Recent analyses suggest a significant divide in how artificial intelligence will impact various professions, distinguishing between digital-centric roles and those requiring physical presence or real-world judgment.

An experiment by researcher Andrej Karpathy used U.S. government data to score 342 occupations on AI exposure. The findings indicate that roles primarily performed on laptops, such as data analysts (score 9) and software developers (score 9), face high automation risks. Conversely, manual trades like roofers (score 1) and plumbers (score 2) are highly resistant. This shift challenges the long-standing ‘learn to code’ career advice, as many programming-based tasks are increasingly replicable by large language models.

Complementing this, a report from Hamilton Group examined 60 occupations with low automation risk. The analysis identified emergency medical technicians and paramedics as having a 0% AI automation risk due to the necessity of real-time judgment and physical intervention in unpredictable environments. The report suggests that while AI may assist with administrative tasks, professions requiring direct human interaction and physical presence remain difficult to replace.