McKinsey: AI May Push 11M US Workers Into New Careers by 2035 | AI Weekly

TEXT START: Roughly 11 million US workers, about 7% of the labor force, may need to switch occupations by 2035 as AI reshapes work, according to a new McKinsey Global Institute report covered by Bloomberg.

The Dissection

The article converts structural displacement into a workforce-mobility problem. It treats 36 million displaced positions and 40 million newly demanded positions as evidence of continuity, while quietly ignoring whether the jobs match in skills, wages, geography, durability, or human necessity. The worker is presented as a movable input, not as someone whose bargaining power may be destroyed.

The headline’s “11 million” is also a containment figure. It counts workers expected to change occupations, not the full population exposed to declining wages, reduced hours, degraded work, or permanent exclusion from economically necessary labor.

The Core Fallacy

It equates aggregate job openings with preserved mass economic participation.

That is the central error under Discontinuity Thesis mechanics. A larger number of openings does not prove that humans remain necessary to produce the output, nor that displaced workers can access the new roles. AI-driven expansion can increase output while concentrating ownership, decision-making, and income among Sovereigns. The remainder are pushed toward narrower Servitor functions or excluded entirely.

“Mobility, not scarcity” is therefore a false dichotomy. The system can have abundant nominal openings and still suffer a collapse in human productive participation. P1 automates cognitive work; P2 prevents institutions from preserving large human-only domains; P3 leaves the majority economically unnecessary. The article measures reassignment before it measures obsolescence.

Hidden Assumptions

• The 40 million new positions will be genuinely human-performed rather than AI-managed, AI-augmented, or later automated.
• Displaced workers possess the time, money, health, education, and geographic freedom required to retrain.
• New occupations will offer comparable wages, status, stability, and benefits.
• Occupational categories are interchangeable enough for “mobility” to solve the mismatch.
• Employers will absorb older, lower-wage, and non-college workers instead of selecting a smaller elite workforce.
• Growth in total demand will translate into broadly distributed income rather than greater returns to AI capital.
• The 2035 labor shortage caused by aging will preserve worker leverage instead of accelerating automation.
• A worker who remains employed remains economically secure. That is not established by the data.

Social Function

This is transition management wrapped in partial truth and ideological anesthetic. The disruption is real, the concentration in vulnerable occupations is real, and the retraining burden is real. But the “more jobs than today” framing functions as a lullaby for policymakers and employers: it relocates responsibility onto workers’ adaptability while leaving ownership and control of the automated productive system unexamined.

It also provides elite self-exoneration. If workers merely need to move faster, then no one has to confront the terminal question: who receives the output when labor is no longer required?

The Verdict

This is not evidence that the mass-employment system survives. It is an early triage memo describing the first evacuation from occupations AI is hollowing out.

The 11 million career switches are the visible layer. Beneath them lies the larger structural threat: workers may find new titles without regaining bargaining power, stable wages, or productive necessity. Retraining can manufacture some Servitors. It cannot restore the mass wage–consumption circuit once AI capital becomes dominant.