The question behind AI panic: What happens to me if it works? - Fortune

TEXT START: Imagine your boss praising a report you finished with AI.

The Dissection

The article identifies the real incentive conflict: workers conceal AI gains because proof of higher productivity can become evidence that fewer workers are needed. Its historical comparison to Taylorism is useful. Its proposed solution—participation, written guarantees, and sharing productivity gains—is a labor-relations intervention designed to preserve cooperation during transition.

But it treats a systemic displacement process as a negotiable workplace bargain. The article sees the fuse and mistakes it for a dispute over compensation.

The Core Fallacy

It assumes AI productivity gains remain a distributable surplus inside a stable employment system. Under the Discontinuity Thesis, AI does not merely make workers faster. It progressively severs the mass employment → wage → consumption circuit.

A six-month job guarantee is a lag defense. Union participation can delay layoffs, improve terms, and redistribute some gains. Neither changes the competitive pressure to reduce labor dependence. A firm that preserves redundant staffing indefinitely risks being undercut by one that does not.

The article also assumes that time saved can be converted into training, backlog reduction, or business expansion that still requires human labor. That may happen temporarily. It cannot be the terminal answer once cognitive automation becomes cheaper, scalable, and broadly superior. The worker’s fear is therefore rational: demonstrating AI competence may be equivalent to demonstrating that the role itself is compressible.

Hidden Assumptions

• Firms can voluntarily share efficiency gains without sacrificing competitive position.
• Jobs remain economically necessary after their tasks are automated.
• Training and redeployment will create enough new human work to absorb displaced labor.
• Employee bargaining power will survive the collapse of labor scarcity.
• Written promises will withstand budget pressure, leadership changes, and competitive shocks.
• Participatory design can coordinate outcomes across firms rather than merely improve one workplace.
• Human oversight will remain indispensable rather than becoming another automatable layer.
• Productivity gains will expand demand for workers instead of increasing output with fewer people.
• The central problem is insufficient trust, rather than the ownership and control of AI capital.

The article acknowledges several of these weaknesses, especially the gap between Norway’s union system and the United States. That honesty makes it a partial truth, not a solution.

Social Function

Classification: partial truth, transition management, and ideological anesthetic.

The text gives workers an accurate explanation for why they hide productivity: the reward structure is punitive. It also offers employers a method for extracting cooperation before automation savings are fully captured. What it does not confront is the terminal question: what happens when the firm no longer needs most of the people whose consent it is soliciting?

Participation can make displacement orderly. It cannot make displaced labor necessary. The proposed bargain is therefore a form of carcass management for the employment system—useful during decomposition, powerless against the underlying death.

The Verdict

The article is a competent diagnosis of worker behavior and a weak theory of system survival. AI panic is not technological confusion. It is accurate perception of obsolescence.

Under DT mechanics, employees who reveal large productivity gains risk accelerating their own substitution, while employees who conceal them merely delay the same process. Norway-style bargaining may buy time and extract concessions, but it cannot defeat P1, P2, or P3. The durable escape routes are ownership and control of AI capital, indispensability in scarce physical and institutional systems, or transition niches around energy, logistics, maintenance, verification, and carcass management.

The article asks how to give workers a reason to share the shortcut. The harder question is what remains for them after the shortcut no longer needs a worker at all.