US-China AI race is heating up, but these 10 professions are still too expensive for robots to replace - is your job safe?

Jobs safe from AI: The United States and China are competing to dominate artificial intelligence, but the biggest challenge for AI may not be building more powerful systems. It may be replacing people who perform the most common jobs.

More Americans believe China is ahead of the US in AI development, as per a new Pew Research Center survey. But a separate study found that several common American jobs could cost far more to automate than to keep staffed by human workers.

The contrast highlights a gap between the global race for AI leadership and the practical limits of automation in the workplace.

Americans are more likely to say China leads the US in AI

A Pew Research Center survey of 3,488 US adults conducted from June 22 to 28, 2026, found that 36% believe China is more advanced than the US in developing AI.

Only 12% said the US is more advanced. Another 18% said the two countries are about equally advanced, while 33% were unsure.

Men were more likely than women to say the US is ahead, at 18% compared with 7%. However, more men and women still said China is more advanced, at 31% and 41%, respectively, as per the study.

Republicans and Republican leaners were more likely than Democrats and Democratic leaners to say the US is leading in AI, at 17% compared with 9%.

AI chatbot users were also more likely than non-users to say the US is ahead, at 16% compared with 9%. But 36% of both groups said China is more advanced.

Does the AI race matter to many Americans

The survey found that 43% of Americans believe it is extremely or very important for the US to be the world's leader in developing AI. Another 22% said it is not too or not at all important.

Men were more likely than women to say US leadership is extremely or very important, at 50% compared with 36%.

Americans with a bachelor's degree or higher were also more likely to say this than those with some college education or less, at 50% compared with 39%.

Older adults were more likely than younger Americans to place high importance on the US leading in AI.

Republicans were also more likely than Democrats to say being the AI leader is important, at 54% compared with 34%.

51% say AI could increase the gap between rich and poor countries

The Pew survey also found widespread concern about the global impact of AI. About 51% of Americans said artificial intelligence will increase the gap between rich and poor countries.

Another 16% said AI would not make much difference, while 7% said it would decrease the gap. A quarter were unsure.

Democrats were more likely than Republicans to say AI will increase the gap, at 60% compared with 45%.

Americans with a bachelor's degree were also more likely to believe the gap between countries would widen, at 62% compared with 46% of those without a bachelor's degree.

Among AI chatbot users, 55% said AI would increase the gap, compared with 47% of those who do not use chatbots.

Why replacing workers is not always as easy as building AI

A June 2026 study by construction scheduling platform Planera examined 30 common jobs in America to determine how much it would cost to replace one worker with automation each year, as per the data shared to The Economic Times in an emailed statement.

The calculation included the cost of purchasing and setting up the technology, operating and maintaining it and paying the human supervisors still needed after automation.

The findings showed that some of the most expensive jobs to automate are not necessarily the highest-paid professions.

Nursing assistants are the most expensive workers to replace

Nursing assistants ranked first. Their median annual wage was $42,200, while the estimated annual cost of replacing one worker with automation was $375,100.

That is nearly nine times the worker's annual wage. The study said the job would require bespoke or humanoid robotics because it involves hands-on patient care. The technology needed does not currently exist as a commercial product.

Physical work is proving difficult for robots

Home health and personal care aides ranked second. The estimated annual cost of automation was $282,200, compared with a median wage of $35,800.

Construction laborers ranked third, with an estimated automation cost of $285,300 compared with a median wage of $47,100. The study said no commercial robot can currently handle the full range of tasks required on a construction site.

General maintenance and repair workers ranked fourth. Replacing one worker was estimated to cost $286,900 a year, compared with a median wage of $49,500.

The job involves diagnosing faults across different equipment types, and the study said no general-purpose repair robot can currently handle all of those tasks.

Teaching assistants and nurses are also difficult to automate

Teaching assistants ranked fifth. Replacing one assistant was estimated to cost $194,300 annually, compared with a median wage of $36,700. The job also involves child supervision and classroom safety responsibilities.

Registered nurses were also included among the 10 most expensive professions to automate. Replacing one nurse was estimated to cost $505,400 a year, compared with a median wage of $97,500.

For elementary school teachers, the estimated annual cost of automation was $213,600, compared with a median wage of $63,900.

Security guards, miscellaneous assemblers and fabricators and restaurant cooks completed the top 10.

Cashiers are cheaper to automate than many other workers

Cashiering was the cheapest of the 30 professions examined to automate. Self-checkout systems were estimated to cost around $24,000 a year, less than the annual wage of a cashier.

The difference shows that automation is not only about whether technology can perform a task. It is also about how much it costs to replace the person doing the job.

White-collar jobs could face greater exposure to AI

A Planera spokesperson told The Economic Times in an emailed statement that the findings challenge the assumption that low-paid, low-skill jobs would be the first to disappear.

The spokesperson said, "The workers earning the least tend to be the ones doing the most physically demanding, human-facing work. And that turns out to be exactly what machines struggle with most. It is white-collar roles that are now more exposed. Software developers, once considered untouchable, are already feeling it. Major tech companies have been cutting engineering headcount precisely because AI coding tools are getting good enough."