Here Are 34 Jobs AI Could Never Take—And How Much They Pay

A worker in a cherry picker and protective gear repairs power lines.

Ariel Skelley/Getty Images

If you want a high-paying job that’s completely safe from being automated by artificial intelligence, you’d better not be afraid of heights or electricity. 

That’s according to researchers at the University of Pennsylvania, who identified 34 occupations that AI could never do in a working paper published last week. These jobs have no tasks that could be done by artificial intelligence.

The top paying position was “Electrical Power-Line Installers and Repairers.” The folks in cherry-pickers who maintain and fix power lines make a median annual salary of $78,310, according to data from the Bureau of Labor Statistics (BLS). The job may be safe from automation, but it is a risky way to make a living—it had the 10th-highest fatality rate of all occupations in 2021, according to BLS data, with 22 deaths reported for every 100,000 workers.

Other relatively high-paying AI-proof jobs included athletes and pile-driver operators. Many, however, had relatively low wages, like bartenders, cafeteria workers and other manual laborers. 

As large language model (LLM) AI products such as Chat GPT grab headlines for their impressive and sometimes bizarre abilities with natural language, tech companies are pouring billions into AI projects

Supporters and critics say AI could soon do many jobs, or at the very least make workers more efficient in these roles, reducing the need for labor. Some of the highest-paying white collar jobs in finance and software are the most exposed, the University of Pennsylvania researchers predicted in their paper. 

Jobs in “Securities, Commodity Contracts, and Other Financial Investments and Related Activities” (average annual wage of $119,040), “Insurance Carriers and Related Activities” ($76,810); and “Data Processing, Hosting, and Related Services” ($97,690) had the most exposure to AI, according to the researchers. 

To calculate the likely AI impact, the researchers themselves—Tyna Eloundou, Sam Manning, Pamela Mishkin, and Daniel Rock, along with other AI experts—rated the individual tasks that each occupation required on how much time could be saved by using AI. The researchers automated some of their own jobs and asked an AI program (GPT-4) to do the same, noting it came up with similar ratings as the humans. 

All told, 15% of tasks across all jobs could be done faster by AI with no decrease in quality, with that number jumping to somewhere between 47% and 56% if the AI were paired with specialized software and tooling for specific applications, the researchers calculated.

“These technologies can have pervasive impacts across a wide swath of occupations in the US,” the researchers wrote. “As capabilities continue to evolve, the impact of LLMs on the economy will likely persist and increase.”

Article Sources
Investopedia requires writers to use primary sources to support their work. These include white papers, government data, original reporting, and interviews with industry experts. We also reference original research from other reputable publishers where appropriate. You can learn more about the standards we follow in producing accurate, unbiased content in our editorial policy.
  1. Arxiv. "GPTs are GPTs: An Early Look at the Labor Market Impact Potential of Large Language Models."

  2. Bureau of Labor Statistics. "Electrical Power-Line Installers and Repairers."

  3. Bureau of Labor Statistics. "Civilian occupations with high fatal work injury rates."

  4. Bureau of Labor Statistics. "Occupational Employment and Wage Statistics."

Take the Next Step to Invest
The offers that appear in this table are from partnerships from which Investopedia receives compensation. This compensation may impact how and where listings appear. Investopedia does not include all offers available in the marketplace.