I saw this going viral, “[Where Is the AI Jobs Crisis?](https://www.apollo.com/wealth/the-daily-spark/where-is-the-ai-jobs-crisis)“:
If AI were triggering a jobs crisis, we would expect job openings to collapse and unemployment to climb, yet the opposite is happening.
The number of job openings per unemployed worker has started to rise again and is now back above 1.0, meaning there are still more jobs available than workers to fill them, see chart below.
The May jobs report reinforced this with nonfarm payrolls jumping by 172,000, confirming that there are no signs of workers being replaced by ChatGPT.
This agrees with earlier posts that the inevitable or awaited “AI job crisis” is a creation of the media and various bloggers and podcasters, who have staked their reputations on this belief that AI will lead to mass joblessness.
If job openings are unconvincing, you can also look at the official unemployment rate, which remains historically low at just 4.3-4.7%. The point is, employers are NOT downsizing due to AI. Labor is one of the biggest costs, so surely if employers could replace expensive workers with chatbots, they would jump at that opportunity. These podcasters and bloggers may eventually be right, but at the very least they should acknowledge that their forecasts have so far been wrong. This is the intellectually honest thing to do.
AI cannot both both be a boon to productivity and result in net job loss. This is unpresented in the history of economics. The “doomer” argument implies that AI becomes super-effective at making employers more productive, and this surplus is not seen anywhere else in the US economy. Macro theory instead predicts that the wealth will be felt elsewhere.
If America becomes wealthier due to AI, it will mean more follow-on jobs such as people using their newfound fortunes to remodel their homes, consume more , vacation, etc. Companies will expand and hire. All of this creates jobs even if AI may also destroy some jobs. The net result is more jobs.
For example, there is a huge market for upper-middle-class people in their 30-50s to look younger. This means more health clinic jobs and demand for pharma. As society becomes wealthier , it means more money spent on elective procedures and healthcare overall. Wellness clinics are a huge deal now. Rich people are outsourcing their healthcare to clinics and other private care, not waiting hours at the “ER”. The biggest users of GLP-1 drugs are not obese people, but the upper-middle class who can afford these drugs to lose the last stubborn 10-20 lbs.
Doordash and Uber Eats are so successful due to the upper-middle-class having so much disposable income to afford these expensive delivery options. Hence, more jobs. Home remodeling is a hue deal ,too. Or landscaping. Go to any expensive neighborhood and you will see construction everywhere, compared to 20-30 years ago when homes were not maintained as well.
A second reason AI has not destroyed white-collar jobs is due to the inherent difficulty of solving coordination problems. The media have a misunderstanding of what these jobs entail. A lot of white-collar work is not about actually “doing stuff”, such as filling out a database, but about organizing and directing.
So you got “party A” and “party B’ and they need to coordinate on “X”. This an be advertising, a legal threat, or answering an invoice. For example, “Company A” wants to buy advertising, so the ad executive talks to the ad representative at “company B”. Onboarding is a major part of Meta and Google. Another example is healthcare. This is involves coordinating many parties, such as:
Doctors
Nurses
Specialists
Patients
Insurers
Off-site testing facilities
Hospitals
Ambulances
Billing departments
Pharmacies
Lawyers (when a claim is denied) Or for car insurance:
Adjusters
Autobody shops
Police
Hospitals
Doctors
Insurers
EMTs (Emergency Medical Technicians) Lawyers
Detectives (if fraud is suspected) Hospitals in both cases outsource their pharmacies, testing facilities, and transportation. This involves more coordination. Often there are endless phone calls and emails between these groups to get things sorted out. It’s hard to imagine how this can be automated. If a single link in this chain is broken, a patient may not get the medicine he or she needs, or a procedure may be delayed, test results lost, etc.