AI革命的政治经济社会学/The Political Economy and Sociology of the Revolution

—

The Political Economy and Sociology of the AI Revolution

Productivity has fissioned, yet relations of production remain stuck in the past

By Xu Daoyi × DeepSeek Civilization Lab

By the time GPT-4 learned to write poetry, Midjourney won design awards, and AI agents began scheduling meetings, writing code, and handling customer service on our behalf — the exclamation marks had barely landed before question marks spread beneath our feet.

At the heart of this intelligence-driven productivity surge, a structural tension is tearing apart the basic framework of modern society: Productivity is undergoing fission-like leaps, while relations of production remain frozen in place. The economic foundation is groaning under strain, yet the superstructure hesitates to take disruptive steps.

This is not another routine technology upgrade. It is a civilization-level restructuring of political economy.

—

01 Productivity Fission: When Intelligence Becomes a Public Good

A genuine revolution never just makes the rich richer — it turns what was once scarce into something cheap or even free.

What the AI revolution is doing, precisely, is transforming “intelligent labor” from the privilege of a few elites into a public good with near-zero marginal cost.

· Efficiency leap: A single AI agent can learn, communicate, and make decisions 24/7, with near-zero iteration costs. What once required a team of twenty for data analysis, customer service, and content creation can now be done by just three to five “human + agent” teams.
· Skill flattening: The once-insurmountable skill gap between junior and senior roles is being filled in. A recent humanities graduate, armed with the right AI tools, can produce insights rivaling those of a seasoned data analyst in just a few hours. The moat of “experience” is draining.
· Organizational evolution: The traditional “company + employee” model is being challenged by project-based collaboration in the form of “individual + AI cluster.” One person as an army is no longer an inspiring slogan — it is a reality.

This is productivity fission — not linear growth, but exponential explosion.

—

02 Frozen Relations of Production: Profits Run Ahead, People Are Left Behind

Yet here is the problem: the energy of fission is not being evenly distributed to everyone.

Most companies today, in their response to the AI wave, still operate under the zero-sum logic of old production relations:

· Profits flow mainly to capital and a small group of core AI owners;
· Middle-skill jobs are being replaced at scale by AI — in some industries, the substitution rate has already exceeded 30%;
· Labor’s share of total income is falling rather than rising. At one major internet platform, after introducing an AI customer service system, labor compensation as a percentage of revenue dropped from 11.2% to 8.9%.

“Frozen” does not mean technology has stalled. It means that distribution rules, organizational models, and labor protection systems have failed to keep pace with the technology.

The result: efficiency gains are captured by a few, while the risks of uncertainty are socialized onto the whole society.

—

03 The Structural Sway of the Economic Base: From Olive to Thumbtack

When a deep rupture opens between the forces and relations of production, the economic base itself cannot remain stable.

Polarization of labor value: Standardized, execution-oriented knowledge work (from translation to basic coding) is depreciating fastest. High-level cognitive creation, complex contextual decision-making, and cross-domain innovation are becoming increasingly expensive. The economy is shifting from a “knowledge-skill premium” to a “cognitive-creation premium.”

Deformation of social structure: Research by sociologist Liang Yucheng shows that repetitive cognitive labor is the first wave of AI-driven displacement. This is precisely the moat on which the traditional middle class has stood. When that moat is drained, the social structure shifts from an “olive shape” (large middle, small extremes) to a “thumbtack shape” (a tiny elite at the top, a vast base at the bottom).

This is not alarmism. History teaches us: the Industrial Revolution gave birth to the working class; the information revolution created the white-collar middle class; the AI revolution is shaping a binary society of “a very few creators + a mass of fragmented task contractors.”

—

04 The Superstructure: Still Turning the Page, While the World Has Changed Books

Every major technological revolution ultimately relies on disruptive institutional innovation to complete the cycle.

After the steam engine, it took nearly 70 years to produce factory acts, legalized trade unions, and compulsory education. This time, the window left for us is being compressed to an extreme.

The dilemma of global governance is one of superficial prosperity but substantive disconnection.

· China: Pursuing context-specific, granular legislation that balances development and security, encouraging innovation alongside regulatory adaptation.
· European Union: Advancing risk prevention through the AI Act, prioritizing ethics and rights protection, though innovation-suppression effects are becoming visible.
· United States: At the federal level, a “development-first” orientation; at the state level, diverse and sometimes fragmented rules — a pattern of “federal looseness, local tension.”

Everyone says “we must keep up with the times,” but no one knows exactly where “keeping up” leads.

In our assessment, the following policy innovations will inevitably come to the table within a few years:

· Impose an AI specific tax on high-profit AI companies and establish a universal data dividend fund;
· Treat core AI infrastructure as critical public resources and explore public or quasi-public ownership;
· Push for a comprehensive revision of labor law that separately guarantees “human-AI collaborative working hours” and “purely human labor hours”;
· Build a Universal Basic Services (UBS) system — not just giving money, but ensuring every citizen access to quality education, healthcare, and retraining.

—

05 The Great Psychological Tension: The Sense of Dread Is Not Exaggeration

Why do so many people feel uneasy or even afraid of AI, even when it brings convenience?

Because AI attacks not only jobs, but also the “meaning of existence.”

For thousands of years, humans have proven their worth through labor. When labor is no longer necessary, or is alienated into “assisting AI,” a deep existential anxiety begins to spread. This is not weakness — it is a civilizational instinct.

At the same time, algorithmic echo chambers, information overload, and digital ecosystems of dubious authenticity further tear apart social consensus. People no longer argue over facts, but over “whose model you are invoking.”

We must face this psychological tension squarely, not mock it.

· Safety bottom lines: data sovereignty, algorithmic explainability, preventing AI attacks on social infrastructure;
· Ethical red lines: no systematic manipulation of human behavior by AI, no deepfakes designed to incite social unrest, no weaponization of AI for autonomous killing;
· Psychological adaptation: build universal AI literacy education, promote new career narratives of “human-AI collaboration,” and help every person find their own ecological niche in the AI age.

—

In Lieu of a Conclusion: Civilization’s Choice Is Being Made Right Now

As Xu Daoyi once said: “Every true leap forward in productive forces has come at the cost of painful restructuring of production relations — but the greatness of civilization lies in its ability to shorten that pain with wisdom and anchor its direction with institutions.”

We stand at a historic inflection point. Will AI become a “super-tool” for a few to dominate the many, or a “common assistant” that benefits all of humanity?

The answer does not lie in technology. It lies in the political choices and social designs we are about to make.

Productivity has already flown out the window. Relations of production are still fixing the door lock.
It is time to tear down the entire wall.

Because the true future is never written in code by technology — it is shaped by the paths we choose together.

—

Xu Daoyi × DeepSeek Civilization Lab
Using the lever of thought to move the course of civilization