A new form of job in the government report

On August 26th, the State Council issued an "Action Plan for Artificial Intelligence Integration".

I specifically downloaded the original document and read it carefully. To be honest, there is a lot of content in the document, covering industries, consumption, governance, and international cooperation.

There are already quite a few interpretations outside. I roughly glanced through them. They are okay, but rather vague. Anyone could write something similar. Honestly, to know what's in the document, running an AI - based summary would suffice.

The sentence I'm most concerned about among so many paragraphs is: "Explore and cultivate the development of innovative work forms such as intelligent agents."

01

Pay attention to the wording, it's "work forms". This means that to some extent, the government has recognized that in the future, jobs will not only be assisted by AI, but many will also be replaced by AI, and some will even be directly undertaken by "intelligent agents".

In the past decade, the common official expressions were "Artificial Intelligence + Industry", "Artificial Intelligence + Education", "Artificial Intelligence + Healthcare" - the core logic was still: AI enabling.

This time is different. It directly puts AI into the labor market, acknowledging that jobs and organizational structures may be rewritten.

This is a strong signal: for the first time at the official level, the government is regarding AI as a formal player in the labor market. In other words, AI may be defined as a "quasi - worker".

If it is really recognized as a work form, it may be included in the enterprise's employment statistics and even affect the calculation methods of social security and taxes.

In an international comparison, this is even more special.

The AI policies of the United States and the European Union mostly emphasize ethics and risks, such as how to prevent the abuse of AI and how to protect employment. However, this Chinese document is doing something else in advance:

We need to actively design a new order. So, this is one of the signals I'm concerned about: the future job market is reserving a place for "intelligent agents".

Why now? Why at this point in 2025? Subjectively, Zhiyuan believes that looking at the development of AI over the past few years on a timeline, we can find three overlapping backgrounds.

First, the technological threshold has truly dropped.

From last year to this year, the progress of large - scale models and agents has been very rapid. In the past, when we talked about AI replacing jobs, it was mostly in the realm of "concept hype".

Now, even small teams can use existing models to build an intelligent agent and make it run in real - world work scenarios.

From writing code and copywriting to helping with market research and process approvals, intelligent agents have shown sufficient stability in some positions.

Take myself as an example:

Recently, I often have to deal with Excel spreadsheets, create some case - closing reports and conduct data statistics. In the past, for these tasks, I had to change column headers line by line, calculate data, and insert links.

Now, I simply throw the spreadsheet to the AI and use a conversational approach to let it handle everything step by step. Finally, it generates a preliminary report, which I can basically use after a little modification.

Another time, a friend asked me to come up with ideas together. I fed the background and ideas to the AI. It first broke some conventional frameworks and then supplemented with new ideas. The result was quite good.

So, in the past, when we discussed "whether AI will replace humans", it was mostly speculation; today, it has become a tangible prototype.

02

Second, the pressure on the employment structure is becoming more and more real.

To be honest, in the past two years, everyone has clearly felt that there are fewer jobs, more people are competing, there are not as many opportunities in big companies, and many traditional positions are shrinking. It's not as stable as before.

Meanwhile, the impact of AI has already begun. For example, among designers, a designer who can use AI may be able to handle the workload of three or five ordinary designers.

The same goes for those in operations and market analysis. People who can use AI are twice as efficient. As a result, a few people become stronger with the help of AI, while a large number of middle - level positions are squeezed out.

The government certainly needs to plan ahead. If we completely let the market evolve on its own, it's very likely that when AI rushes in, a batch of jobs will disappear quickly, and many people will be forced to change jobs or even become unemployed.

So, the document simply describes "intelligent agents" as a "work form". The underlying meaning is:

Don't simply see it as "job losses". We need to restructure it and turn it into a new form of labor.

To put it bluntly, this is like giving society an early "vaccination". Instead of being passively beaten, we should directly recognize that AI will change the labor market and then incorporate it into a new framework.

Speaking of this, we have to take a look at the outside world.

In this wave of AI, everyone is vying for a position. The focus of discussions in the United States is more on risks, so they emphasize legislation and ethics. The European Union is even more strict, directly introducing the "AI Act" to impose strict restrictions on high - risk scenarios.

To put it simply, they step on the "brakes" first, worried about losing control of the situation. China, on the other hand, chooses a path of active definition: first recognize that AI will rewrite the labor market, and then write the rules in advance.

This is a form of institutional competition. Whoever defines the new order of the labor market first will have more say. In the future, in addition to competing in technology and computing power, we will also compete in whose institutions can better accommodate this new species of "intelligent agents".

Putting these three points together, we can roughly understand why this sentence was included in the State Council's document in 2025.

03

If "intelligent agents" really become a new work form, what specific impacts will it bring? Zhiyuan believes we can look at it from three aspects:

First, for industries.

In the past, when we talked about "human - machine collaboration", it mainly stayed at the tool level.

To put it bluntly, it was about using a software to improve efficiency or a system to automate some processes. If "intelligent agents" can really replace jobs, the division of roles in enterprises will be completely rewritten.

Let me give a recent example:

In the enterprise service field, DingTalk, Feishu, and WeCom for Enterprises are in fierce competition. In the past, they were more "tool - oriented", helping you streamline processes in areas such as schedules, meetings, spreadsheets, and approvals.

Now, the evolution is as follows: First, everyone adds AI to existing tools to achieve more efficient automation. For example, meeting minutes can be automatically generated, and spreadsheets can be automatically calculated.

Second, after adding AI, a layer of "Agent" is added. It's like a chat window where you can directly give instructions: arrange a schedule for me, organize meeting minutes, create a to - do list.

At this time, the collaboration model changes: humans and intelligent agents work together. To put it simply, it's a new "hybrid model".

Imagine that in the past, an assistant would help you arrange schedules, take meeting minutes, and follow up on tasks. Now, an Agent can handle it all. Moreover, it can integrate data and processes across different tools automatically.

Over time, in a team, in addition to human - to - human cooperation, people will also habitually regard agents as "virtual colleagues".

Looking further ahead, "AI - native enterprises" may emerge. What does this mean?

From the very beginning, an enterprise does not have a complete human - resource structure but is based on "human - machine hybrid teams" as the basic unit. For example, a startup with a founder and five or six intelligent agents can achieve a business scale that used to require dozens of people.

This will bring two obvious changes:

One, the improvement in efficiency will far exceed expectations. Two, industry boundaries may be redefined. Many barriers that used to rely on human resources will be broken due to the popularization of agents. At that time, enterprises will compete in the combination ability of "talent + number of agents".

The change at the industry level is the rewriting of the way companies operate, and at the individual level, jobs may completely change.

Previously, we always thought that one person held one position. After the emergence of "intelligent agents", the concept of "super - individuals" emerges.

What is a super - individual? One person + N agents, combined into a more powerful unit.

For example, a designer used to have to stay up late to revise ten versions of a design. Now, he only needs to set the tone, clarify the style, let several agents produce drafts respectively, and then select the best one for modification.

This trend will lead to a situation where those who can use agents will be extremely powerful, while those who can't will soon be marginalized. In other words, people will become stronger with the help of agents.

04

As industries and individuals change, the institutional level will surely change as well. Although we haven't seen the specific design yet, the problems are already there:

Will the social security system be adjusted? In the future, when calculating pensions and medical insurance, can we still simply count by the number of people, or do we need to consider the combination of "human + agent"?

Don't forget, the document not only mentions employment but also scenarios such as "intelligent companionship" and "health assistants". When jobs are redefined, should our concepts also change?

For example:

In the past, "working hard" meant personally completing a certain number of tasks. In the AI era, will it become "how many agents you have and how many achievements you have produced"? Should we also re - examine some traditional values that were emphasized in the past?

The sentence in the document, "Promote the transformation of research methods in philosophy and social sciences towards human - machine collaboration", actually reminds us:

The future challenge is that our ideological understanding will undergo a new transformation, and even a brand - new education method may emerge.

I don't have answers to these questions. But I can see that it's hard to say whether it's good or bad, and perhaps it can't be measured by "good" or "bad" at all. What's certain is that we must learn to collaborate with AI and upgrade ourselves into "super - individuals".

I'm not spreading anxiety. When we discuss "intelligent agents", it's more like a mirror reflecting a corner of the future society. A society where humans and AI coexist may be quietly approaching.

Of course, there's no need to be overly pessimistic.

After all, every time a technological wave arrives, we can always find a new order in the chaos. The key is that those who can quickly learn to collaborate with "intelligent agents" and upgrade themselves into "super - individuals" will be the first to gain a foothold in the new order.

This article is from the WeChat official account "Wang Zhiyuan" (ID: Z201440), author: Wang Zhiyuan, published by 36Kr with authorization.