As Labor Day has just passed, I thought it might be interesting to reflect on the current impact of artificial intelligence on the labor market. Is a backlash against AI taking shape? I vividly remember the audible gasps at a conference that I attended last year when the technology entrepreneur Vinod Khosla (who also just purchased the Seattle Seahawks football team) said that AI would replace 80% of the work for 80% of all jobs.

This revelation by an industry oracle was followed up by Anthropic founder Dario Amodei’s dire prediction that AI would wipe out half of white-collar jobs. Further, as Noah Smith recently pointed out in his Substack column, it is an article of faith in the San Francisco technology community that artificial intelligence will lead to significant job losses. Surveys show that employees are very nervous about their employment, and for good reason because while AI offers the promise of great scientific advances, perhaps its primary business objective is to replace the labor of costly human beings with less expensive technology that works 24-7.

However, the temporal good news is that the sky has not fallen – yet. According to a recent article in the Economist entitled “The jobs apocalypse is postponed. An AI jobs boom is here”. It reported that AI has led to the creation of an additional estimated 1 million jobs in America while only 200,000 layoffs have been attributed to AI since mid-2023. For those of you keeping score at home, that’s a net gain of 800,000 jobs.

Torsten Slok, chief economist at Apollo, came to a similar conclusion that AI is retraining workers, not replacing them. As evidence of his thesis, Slok cites an August survey by the New York Federal Reserve Bank of companies in the New York and Northern New Jersey region which showed that 34% of service businesses and 22% of manufacturers that use AI are retraining their staff, while only 4% and 0% respectively report layoffs. Accordingly, the Fed report asserts that businesses are using AI to transform work and not cut jobs.

As a corollary, Slok notes that his own research shows that AI has created downward pressure on wages in AI-exposed occupations while employment levels in these occupations remain unchanged. In his view, that suggests that companies are reaping the benefit of productivity gains from AI through wage compression instead of workforce reduction.

But we shouldn’t get too comfortable with the idea that AI will not result in a net loss of jobs and that it won’t completely disrupt society. Bill Gates told Semafor that “there isn’t a job that isn’t affected by AI” and he predicted that in the majority of occupations, humans will be “vastly inferior to the AI doing the job.” Yikes! Gates is also concerned that the backlash against AI due to job losses will be so great that “It’ll make the data center thing look like nothing.” He fears that if letting AI continued unchecked will lead society to turn against it entirely, thus derailing advancements in science and health which will improve humanity.

Now let’s come back to what we at Wharton Property Advisors do, which is to help tenants get New York City office space. I always like to ask this relevant question: what does this have to do with commercial real estate? Once again, the answer is a great deal (which is perhaps a double entendre). That is because companies’ office space needs are connected to employment, and if there is continued growth, that is good for our industry. In addition, one of the major reasons for the strength of the NYC office market, and the principal reason for the uptick in the San Francisco market is due to the additional office space taken by AI companies due to needed expansion along with a positive changes instituted by that city’s relatively new mayor. Unlike the potential harm caused by data centers, this growth is good for our urban economy.

This is also a cautionary tale about the dangers of following the conventional wisdom. For years, I have argued that the conventional wisdom should be disdained as it is often wrong. Let’s hope that is the case here, or otherwise we will have a massive problem of unemployment on our hands. This is truly a case where we can hope for the best, but we need to prepare for the worst, which is why experts like Gates have suggested solutions like designating certain occupations as “human reserved” even if machines do them more efficiently and taxing AI computer tokens and robots.

Although it is just the early innings of AI, the doomsday job scenario has not come to pass for the time being. However, you don’t have to be Hamlet to know that the question of whether AI will create or destroy jobs remain remains unclear while outrageous fortunes are being made by its founders. The answer is muddled as there have always been winners and losers in times of dramatic technological change. We’ll just have to wait to see what happens.

Now here’s a final thought. It’s somewhat of a paradox that the all-conquering AI on which literally trillions of dollars are being spent is also considered to be so harmful to education that New York City has just banned all student use of AI in public schools through eighth grade and has significantly limited its use in high school for one year, with sensible exceptions for students with disabilities and those learning English. Similarly, Los Angeles has prohibited student access to generative AI tools on all district-issued devices for the 2026–27 school year. The reasons for the change make sense to me. First, test scores are down in reading comprehension. Second, teachers have detected a diminution in the critical thinking skills of their students who take the easy way out using AI.

And now there are changes afoot at the upper end of the education system as well. The Financial Times reports that the University of Chicago Law School where President Obama taught is banning phones, tablets, and laptops from first-year classrooms. Apparently, AI is interfering with the time-honored Socratic method of the ancient Greeks which requires law students to think on their feet, a skill that will be essential during their future careers. Similarly, the University of California School of Law is barring the use of generative AI tools for pretty much everything as well. This begs another question: could there be a counterrevolution against the AI revolution afoot beyond the rebellion against data centers?

If AI is so great, why is it being banned in schools and not businesses for non-financial analysis? Perhaps it is because businesses don’t care if the end product is plagiarized if it is better or less time-consuming than the work that can be done by humans. However, business should care if its employees can think clearly and communicate effectively. In addition, what does that say about us, and what does it say about the value of AI that there is a growing consensus that it does not belong in schools because it stunts the intellectual development of students. That growth should continue throughout their entire education, business careers, and retirement years. AI is obviously a valuable tool, but perhaps we run the risk of overusing it at our peril for more than safety reasons. Contrarians take note.

Thank you,
Ruth Colp-Haber