The revolutionary effect that major tech leaders predict artificial intelligence will likely have on jobs is a largely nebulous assertion, industry experts said.
In recent months, prominent tech figures ranging from xAI’s Elon Musk to OpenAI’s Sam Altman have suggested that in a “benign” scenario, AI will make most jobs “not real work,” lead to a society where labor is “optional,” and create a world where a universal basic or high income is possible. The concepts appear to be built on the premise that technology will be able to automate tasks on a mass scale, boosting productivity to unprecedented levels and creating a future of “generally sustainable abundance,” as Musk put it during a podcast with Joe Rogan in the fall.
However, field specialists told the Washington Examiner that major “gaps” remain in terms of how such an AI could come to fruition. Such predictions are largely dependent on the still-theoretical concept of artificial general intelligence, or AGI, they said. And they are also contingent on many “unanswered questions” about how highly developed, AI-powered robotics can perform all human tasks.
“There are gaps in this utopian picture that don’t necessarily connect and don’t necessarily have a clear timeline or a clear causal chain, because it’s not clear, for example, who the robots are working for,” said Kevin Frazier, an AI innovation and law fellow at the University of Texas School of Law.
“It’s not clear who is responsible for powering the robots,” he added. “It’s not clear who’s responsible for maintaining the robots. And so the second you start to ask some additional questions about who the resources would be compiling to, who will be managing both the AI systems, the computer systems, as well as the physical infrastructure that’s required to sustain this vision — these, to me, are among the many unanswered questions.”
Daniel Cochrane, a senior research associate at the Heritage Foundation’s Center for Technology and the Human Person, raised questions about the “high-level rhetoric” about the “really vague notion” that AGI can do “everything that a human being could do, and therefore replace all humans.”
“There’s a lot of smoke and mirrors,” he said. “There’s a lot of disagreement as to what it actually looks like in practice.”
This doesn’t mean AI won’t shake up the economy. It has replaced thousands of jobs in the United States this year alone. And a new study from the Massachusetts Institute of Technology found that it already holds the capacity to replace 11.7% of the labor market.
Cochrane conceded there have been “significant improvements” in AI technology, pointing to Anthropic’s release of its Claude Opus 4.5 model in late November. In a sort of “horseshoe effect,” companies such as Amazon, Target, and UPS are automating low-level jobs, while white-collar jobs in the tech industry and beyond are also being outsourced to technology, he said.
“I think you’re going to see a lot more layoffs, because the water holes in, you know, again, at the mid-tier level, but also the lower tiers in you know, things like logistics and manufacturing are going to get automated away,” Cochrane predicted. “I think there’s certainly the case that we will see significant automation of certain sectors.”
While large companies, such as Walmart, are researching how to harness AI to automate their processes, the technology could also yield significant benefits for smaller businesses and entrepreneurs, Cochrane said. This is because, in a positive scenario, they could use it to reduce operational and labor costs while also channeling it to accomplish time-consuming administrative tasks, allowing them to specialize in a particular area of expertise.
Frazier predicted that in 2026, AI will continue to take over routine tasks in “knowledge industries, such as legal, government, accounting, and human relations professions.” There will likely be a decrease in the overall employment in those sectors, in addition to higher rates of unemployment “among entry-level folks who are looking for jobs” in such arenas, he said.
“There are a lot of specific tasks of those jobs where automation and the use of artificial intelligence are going to reduce the total number of workers that are needed in those industries,” Frazier said. “That’s not to say that AI is going to take over the whole task of lawyering, or the whole task of being an accountant, but by virtue of taking over some of the core tasks that those individuals perform, those firms are just going to need fewer and fewer accountants, lawyers, HR professionals, for example.”
Regarding how laid-off workers could be reintegrated into the labor force, expanding private-public collaboration could offset the negative effects of AI, according to Frazier. Upskilling and retraining programs traditionally spearheaded by only the government have not typically resulted in “long-term employment that’s going to drastically improve their economic security,” he said. That includes a combination of information sharing about how companies are adopting AI and to what effect, as well as creating more apprenticeship programs oriented toward areas where the workforce is growing.
“We need to see that retraining programs are having more ongoing and in-depth conversations with private firms to better anticipate what the sorts of skills are that are going to be compensated and demanded, not only for the next year or two, but looking forward to the next five years, or even the next decade,” Frazier noted. “Having a chance to learn from folks on the ground already doing the work in a way that makes it far more likely that they’re going to gain those skills in a faster fashion and then have increased odds of earning a long-term position.”
Cochrane also said he does not foresee a world that is “going to see an AI that can literally replace all” jobs. Even if it was “technically possible” to achieve full-scale automation, he said he believes there “would be enough friction” in democratic Western governments to prevent such an outcome from occurring.
He argued that society places immense value on human connection, particularly in arenas associated with healthcare and spirituality, saying, “There’s a certain sense of dignity that is lost if there’s no human being in the loop overseeing those machines. And if you think about like a hospital setting, right, you know an elderly patient in a hospital or old folks home or elder care facility, you want a human being that’s caring for them, right?”
Or if you think about a different profession, a counselor or a therapist, or even a pastor, right? These are professions where we recognize technology may shape and change them, and already is, but there’s a sense in which the essential humanness can never be replaced,” he continued.
A YouGov poll conducted last month found that 72% of U.S. adults worried about the broader economic effects of AI. This was similar to a Reuters poll conducted over the summer, which found that 71% of respondents said they were concerned that AI would be “putting too many people out of work permanently.” In that survey, two-thirds of respondents also said they were worried that people would abandon relationships with others in favor of AI companions, echoing 2025 data from the Pew Research Center, which found that 57% of the public was “highly concerned” about AI leading to a decrease in connection between people.
TRUMP SIGNS EXECUTIVE ORDER LIMITING STATE REGULATION OF AI
Cochrane warned that certain “tech elites” are seeking to make the case that AI will “certainly” be able to develop to the point where it can build trust in relationships like a human can. He said that technology is simply doing “what our biology is doing — just better.” However, he added that doing so would be making a “huge metaphysical assumption” about what it means to be humans, reducing people to “biological calculators.”
“When you recognize beauty, or you love another human, that’s not just a matter of, you know, data points being calculated in your head, or a bunch of chemical reactions,” Cochrane said. “Human beings are, by our nature, not just rational calculators. We are souls, and we have the ability to make moral judgments. We have the ability to have experiences that are not necessarily empirical. There are certain human elements that can never be wholly substituted by AI.”
