Low-wage cities such as Las Vegas, Orlando, and El Paso will be hit the hardest by job automation, according to a recent report from the Institute for Spatial Economic Analysis (ISEA). And the impact of automation on job losses is likely to be more severe than previously predicted, the report stated: Due to advances in machine learning and mobile robotics, jobs such as truck driving, healthcare diagnostics, and education are more likely to be affected.

“The replacement of jobs by machines has been happening continuously since the Industrial Revolution, but it’s expected to significantly accelerate in the coming 10 or 20 years,” said Johannes Moenius, founding director of ISEA, in a press release. “Pretty much everyone will be affected, but some metropolitan areas will see a lot more jobs vanish than others.”

ISEA examined Oxford University research on the probability of automation for a number of occupations, as well as employment data from the Bureau of Labor Statistics. Of the 100 metropolitan areas in the US with more than 250,000 people employed, the following 10 cities have the largest share of jobs that may become automated:

1. Las Vegas-Henderson-Paradise, NV

65.2% of jobs automatable

2. El Paso, TX

63.9% of jobs automatable

3. Riverside-San Bernardino-Ontario, CA

62.6% of jobs automatable

4. Greensboro-High Point, NC

62.5% of jobs automatable

SEE: Future jobs: How humans and robots will complement each other

5. North Port-Sarasota-Bradenton, FL

62.4% of jobs automatable

6. Bakersfield, CA

62.4% of jobs automatable

7. Orlando-Kissimmee-Sanford, FL

61.8% of jobs automatable

8. Fresno, CA

61.5% of jobs automatable

9. Greenville-Anderson-Mauldin, SC

61.3% of jobs automatable

10. Louisville/Jefferson County, KY-IN

61.3% of jobs automatable

Almost all large metropolitan areas in the US could lose more than 55% of their current jobs due to automation, the report stated. High-tech hubs such as Silicon Valley and Boston are least likely to be affected.

SEE: Video: Panera’s automation investment is actually creating jobs

At-risk occupations include office and administrative support occupations, food preparation and serving related occupations, and sales and related occupations. These three categories account for half of the automation potential in the largest metro areas. Meanwhile, transportation and material moving positions contribute to potential employment losses in Riverside, Louisville, and Greensboro.

However, the probability of automation does not equal future unemployment rates, said ISEA faculty fellow and report co-author Jess Chen. A recent report from Forrester Research estimated that automation and robotics will displace 24.7 million US jobs by 2027–but that the technology will create 14.9 million new jobs in the same time period, leading to a net loss of 9.8 million jobs.

“Technical feasibility does not imply that automation necessarily makes economic sense. And historically, automation went hand in hand with new job creation both in skilled and less skilled labor,” Chen said in the release. “However, the speed and the high share of automation in less skilled jobs raises many questions about whether the economy will be able to make up for the expected job losses. What we do expect is that automation will create winners and losers among cities and regions of the U.S., where losers may not recover to their original employment levels within even a decade’s time.”