TL;DR
Get everyday essentials delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
A November 10, 2026, Thorsten Meyer AI essay argues that knowledge work has become a major part of modern economies and is now a central target for generative AI. It cites historical estimates and current occupational data, but says AI exposure does not establish job losses and that broad employment data shows no clear AI effect so far.
Thorsten Meyer AI published an essay on November 10 examining how generative AI could affect knowledge work, the large category of jobs built around producing, processing and sharing information. Drawing on economic research dating to the late 1950s, the article argues that AI is directed at work already central to advanced economies, while warning that exposure to the technology is not the same as job loss.
The essay’s historical starting point is economist Fritz Machlup’s 1962 study, which estimated that the production and distribution of knowledge accounted for $136.4 billion, or “almost 29 per cent of adjusted GNP,” in the United States in 1958. The figure measured spending on knowledge-related activity, not the economic value or quality of knowledge itself. The article says education made up the largest share of Machlup’s total, followed by media, information services, research and development, and information machines.
The piece also traces the idea of knowledge work to management writer Peter Drucker, who used the term in 1959. It describes a broader shift from defining work by physical production to including tasks such as planning, writing, calculating, checking and communicating. Citing estimates attributed to Marc Porat, the essay says information workers rose from less than 6% of the US workforce in 1860 to 46% in 1970, becoming the largest occupational group by the mid-1950s.
For a present-day comparison, the article cites 2025 International Labour Organization figures showing managers, professionals and technicians accounted for 62% of jobs in Singapore, 52.5% in the United States and 47.9% in Germany. It says higher-income and clerical roles are among those most exposed to generative AI, but gives no specific estimate of how many jobs will be eliminated or changed.
What AI Means for Knowledge Work in the Fourth Sector
Generative AI targets tasks central to office and professional work. But exposure to AI is not proof of job loss, and broad US employment data shows no clear AI effect so far.
What counts as knowledge work?
The Fourth Sector covers work centered on producing, processing and sharing information.
Teaching and learning
Education was the largest category in Machlup’s historical estimate.
Building knowledge
Research and development were among the activities counted.
Moving ideas
Media, information services and information machines also featured.
Applying expertise
Planning, writing, calculating, checking and communicating are part of the shift beyond physical production.
A long shift, not a direct comparison
Historical estimates and modern job classifications describe different categories and periods.
Exposure is not displacement
AI may assist with tasks within a role without eliminating the whole job.
- Higher-income and clerical roles are among those more exposed.
- Generative AI can handle tasks such as drafting, summarizing, analysis and information processing.
- US employment data shows no clear overall AI footprint so far.
- How many jobs may change or be eliminated.
- Whether exposure leads to changes in hiring, wages or job duties.
- Whether the reported lag among young workers in exposed occupations was caused by AI.
What readers and employers should watch
The key evidence will come from how organizations adopt AI and what changes follow.
Look for evidence of how employers use generative AI in work.
Follow hiring, responsibilities and wages over time.
Compare workers and occupations over time to distinguish AI effects from other economic changes.
AI Reaches the Information Economy
The development matters because generative AI tools are designed to handle tasks that form part of many office, professional and clerical jobs: drafting, summarizing, analysis and information processing. If these tools change how such work is done, the effects could extend well beyond technology firms, touching employers and workers across sectors where information is a principal input.
But the essay’s central distinction is between exposure and displacement. A job may include tasks AI can assist with without the whole role disappearing. The article says overall US employment data does not yet show a clear AI footprint, while reporting that young workers in the most exposed occupations are falling behind their peers. That observation signals a possible area of concern, not proof that AI caused the difference.
For readers, the key question is therefore not simply whether AI can perform particular tasks. It is how organizations adopt it, whether workers’ responsibilities change, and whether those changes affect hiring, wages or career entry. The source offers historical framing and selected indicators, rather than a forecast that answers those questions.
Products Worth Considering
Bose Smart Soundbar 700 with Alexa, Black
LG UltraGear 32GX850A-B 32-inch 4K UHD OLED Gaming Monitor
Amazon Kindle Scribe (3rd Generation), 64GB, 11-inch Display, Graphite, with Premium Pen
As an affiliate, we earn on qualifying purchases.
How Knowledge Work Was Counted
In The Landmarks of Tomorrow (1959), Drucker described the growing importance of workers who earn a living through specialized knowledge. His argument, as summarized in the essay, was that management approaches developed for factories and assembly lines would not automatically suit work whose main inputs are expertise, judgment and information.
Machlup’s later attempt to quantify the trend used spending as a proxy because, as the essay explains, knowledge itself has no physical weight and is difficult to count directly. His five categories included education, research and development, communications media, information machines and information services. He also examined occupations involved in producing and passing on information. The resulting 29% figure refers to adjusted US GNP in 1958; it is not a current measure and should not be read as the share of today’s economy.
The essay connects those early accounts to more recent workforce figures, but the measures are not identical: historical estimates of information or knowledge occupations and modern occupational classifications cover different categories and periods. They illustrate a long-term shift in work, rather than a single directly comparable time series.
“Almost 29 per cent of adjusted GNP.”
— Fritz Machlup, as quoted in the essay’s account of his 1962 study
What Job Effects Are Established
The essay does not specify the studies or methodology behind its statement that higher-income and clerical jobs are at the front of the line for generative AI exposure. It also does not give a definition of exposure or identify the occupations included in that category, so readers cannot infer a precise share of jobs at risk from the information provided.
It says US employment data has shown no clear overall AI footprint so far, but does not name the data series or period used for that conclusion. The reported lag among young workers in highly exposed jobs is also not, on its own, evidence that AI caused the difference. The scale, duration and cause of that pattern remain unclear in the source material.
Tracking Adoption and Employment
The next evidence to watch is whether employers’ use of generative AI translates into measurable changes in hiring, job duties, wages or productivity. Longer-running employment data and studies that compare workers and occupations over time will help distinguish AI-related effects from other changes in the economy.
The essay identifies a follow-up topic in its series: how economic statistics measure activity and what those measurements miss. For the present AI question, no specific forecast, policy proposal or next data release is provided. Readers should treat the article as a historical and present-day framing of the issue, not a definitive account of future job losses.
Key Questions
What is the Fourth Sector in the essay?
It refers to knowledge and information work: activity centered on producing, processing and distributing information, including work in education, research, media, information services and related occupations.
What did Machlup’s 29% estimate measure?
It measured spending on producing and distributing knowledge in the United States in 1958. The essay reports $136.4 billion, equal to almost 29% of adjusted GNP. It was not a measure of knowledge’s quality or total economic value.
Does the essay say AI is already eliminating knowledge jobs?
No. It says some roles are exposed to generative AI but stresses that exposure does not prove job loss. It also says overall US job data has not yet shown a clear AI footprint.
Which workers does the essay identify as more exposed?
It points to higher-income and clerical jobs as being at the front of the line for generative AI exposure. It does not provide a full occupation list or a quantified estimate of potential job losses.
Source: Thorsten Meyer AI
