The General Office of the State Council issues a document requiring a prudent approach to the pace of AI application in labor-intensive industries

[[**An October 23 document from the General Office of the State Council requires a steady and prudent grasp of the pace of AI application in labor-intensive industries, attempting to slow the disappearance of old jobs and accelerate the creation of new ones; but the employment shock, the difficulty young people face in entering the workforce, and the outflow of efficiency constitute trade-offs.**

**Key points**

**1. The core of the document is steady and prudent slowing, not a simple positive**

State Council Office Document No. 29 has 6 aspects and 18 tasks; Article 6 requires a steady and prudent grasp of the scope, intensity, and pace of AI application in labor-intensive industries.

**2. AI employment assessment and job-transfer buffers are written into policy**

**3. Worst-case scenarios are incorporated into large-scale unemployment contingency plans**

**4. AI substitution and youth employment pressure are emerging, and excessive braking also has costs**

After Laoganma's AI quality inspection, the pepper-sorting line dropped from more than 600 people to 48; the graduating class of 2026 totals 12.7 million, and employment in AI-exposed positions among 22-25 year-olds is 19% lower than that of their peers; too-slow application also causes some work to flow abroad. 2026-10-26 08:17 The State Council sets the tone on AI employment: labor-intensive industries must be "steady and prudent" TOP Innovation Zone Research Institute quick overview

This article comes from the WeChat public account: TOP Innovation Zone Research Institute, author: AI Research Group

Recently, the forecasting agency Polymarket posted a news flash that got 500,000 views:

The top reply was this:

At least they're trying to do something; the rest of us can only rot, while being told a new golden age has arrived.

This news flash refers to the "Opinions on Adapting to the Development of Artificial Intelligence Technology and Helping Build an Employment-Friendly Development Model" issued by our State Council General Office on October 23. Securities Times gave it the headline "After the Bell, Good News Is Here!"

On one side is the alarm; on the other, the good news.

We looked at this document, and neither side really mentioned Article 6:

"Steadily and prudently grasp the scope, intensity, and pace of artificial intelligence application in labor-intensive industries."

In 1589, in the village of Calverton in Nottinghamshire, a clergyman named William Lee built the world's first knitting machine.

Lee took the machine to London, hoping to obtain a patent from Queen Elizabeth I.

The Queen looked at it and did not approve it.

According to the most widely circulated version, this is what she said:

Mr. Lee, your ambition is not small, but think about what this thing will do to my poor subjects. It will take away their livelihoods and make them beggars.

Lee did not give up. When James succeeded to the throne, he applied again, and was still refused.

Around 1608, Lee took his brother, 9 workers, and 9 machines to Rouen, France, and King Henry IV of France readily granted him a patent.

Elizabeth was probably the first ruler in history to conduct an "employment assessment" of a new technology.

You may think she was conservative, but in that era, this was actually standard practice for rulers.

Oxford University economist Carl Benedikt Frey specifically reviewed this in "The Technology Trap":

Before the Industrial Revolution, technologies that took away jobs were often held back by rulers; after all, too much unemployment would cause trouble.

More than two hundred years later, the British government showed a different face and stood with the machines.

Starting in 1811, workers in central and northern England smashed machines in groups; this was the famous "Luddite movement." Parliament passed the Destruction of Stocking Frames Act in February 1812, making machine-smashing punishable by death; a total of 12,000 soldiers were sent to suppress it, and historian Hobsbawm said this was more troops than Wellington had at the time against Napoleon in the Peninsular War.

But there was one stage in this process that was especially hard to endure.

Oxford economic historian Robert Allen gave this period a special name: the "Engels' Pause." Engels wrote "The Condition of the Working Class in England" in 1845 about the workers of these years.

According to estimates cited by Allen, from 1780 to 1840, output per worker in Britain rose 46%, while workers' real wages barely moved. Only after 1840 did wages catch up: over the next 60 years, per capita output rose 90%, and real wages rose 123%.

Suppose a technology eliminates 100 jobs in ten years and creates 120 new jobs, for a net gain of 20. But if those 100 jobs disappear in the first two years, and the 120 new jobs only gradually emerge in the eighth or tenth year, then the years in between are a smaller version of the "Engels' Pause."

Elizabeth simply pulled the battery out of the first clock, and as a result the machine went to France;

In 1812, Britain wound the first clock to its fastest, and in exchange got half a century of wage stagnation, plus a suppression force larger than the one on the Peninsular battlefield.

State Council Office Document No. 29 has 6 aspects and 18 tasks, most of them a familiar formula:

The original text is that for artificial intelligence application scenario projects in market access areas that may cause a relatively large employment shock, "an employment risk assessment shall be carried out simultaneously during review."

Everyone has long been used to environmental impact assessments before chemical plants go into operation. In the future, when some AI projects are launched, an employment report will also have to be placed on the review table. What Elizabeth decided alone more than four hundred years ago now has a formal process; the difference is that this time there is no intention of a one-vote veto.

2. Leave workers reaction time: an unemployment buffer period.

One more thing: the origin of this idea is actually quite interesting—at the 2025 National Two Sessions, iFlytek Chairman Liu Qingfeng proposed setting up a 6-to-12-month buffer period, paired with an "AI unemployment insurance." Nineteen months later, this idea entered the State Council Office document.

3. Provide a fallback for the worst case: prepare contingency plans

More than 400 companies are gathered in 3 square kilometers, over 70% of them do AI, the average age of core founders is 28, and "one-person companies" in which one or two people lead a group of "AI employees" are becoming more and more common.

At this autumn's Guizhou Provincial Quality Development Conference, Li Xin, executive vice president and chief quality officer of Laoganma, reported a string of impressive figures:

After AI visual quality inspection went online, the pepper-sorting assembly line at the Changming base, which originally had more than 600 people, retained only 48 people for the final check; revenue in 2025 was 5.4 billion yuan, a record high.

His original intent was probably to show off the company's strengths.

As a result, in mid-October, "79-year-old Laoganma founder Tao Huabi makes 5.4 billion yuan a year with AI" topped Weibo's hot search list, with more than 48 million views, and netizens kept asking the same question:

Where did those 552 people go?

The company did not say.

A comparison with Document No. 29 makes it clear. Pepper sorting is a typical labor-intensive position, and Guizhou is in the central and western region. The phrase "advance notice and reassignment support" in the document is exactly what is missing from this set of figures.

However, on that Laoganma assembly line, it was incumbent workers who were squeezed out by machines,

The greater pressure falls on those who have not yet entered the door.

The number of college graduates in the class of 2026 is 12.7 million, a record high. Opportunities for those aged 16 to 24 are even fewer, while during the same period, the employment rate for those aged 30 to 59 has remained almost unchanged.

Middle-aged people still hold on to their rice bowls, while young people have not even gotten through the door.

Data from the United States is similar.

Stanford Digital Economy Lab used U.S. payroll data to track and found that young people aged 22 to 25 in jobs with high AI exposure had 19% less employment than projected based on the trajectories of their less-exposed peers, with the gap mainly coming from companies hiring fewer people.

On the same day, October 23, Ron Wyden, the top Democrat on the U.S. Senate Finance Committee, unveiled the Worker Stabilization and Opportunity Act: people who lose jobs due to technological change could receive up to 26 additional weeks of benefits, gig workers could also receive job-search allowances, and a "transition wage" subsidy would be set up for retraining.

Where would the money come from?

A consumption tax on data centers.

In the U.S. context, in the first nine months of this year, of the layoffs announced by U.S. employers, about 120,000 jobs listed AI as the reason, ranking first among all reasons and more than double the full-year figure last year.

The U.S. approach is more like letting the first clock run first and making up for lost jobs with money afterward;

China's document is thinking about slowing the first clock down a bit first.

But slowing it down also comes at a cost.

Sound familiar? More than four hundred years ago, William Lee took nine machines to France in exactly this way.

If the brakes are applied too hard, jobs may not be retained and may very likely just emerge in another country.

In the end, it comes down to trade-offs. AI industry signal channel: Frontier Technology. This content comes from the internet, original link; the views represent only the author and do not represent Huxiu's position. If there are copyright issues, please contact hezuo@huxiu.com, and we will verify and handle them promptly. Everyone who is changing and wants to change the world is on the Huxiu APP[ ]