EconomyAI Research

The Reality of AI & Jobs: Moving Beyond the Hype

Based on the Anthropic Labor Market Report (March 5, 2026)

There is a massive difference between what AI could do and what it is actually doing in our offices today[cite: 14, 19]. While headlines often scream about total job replacement, a landmark study from Anthropic titled "Labor market impacts of AI: A new measure and early evidence" provides a much-needed reality check[cite: 2, 7, 70].

The "Capability Gap": AI is far from reaching its theoretical limits[cite: 14]. In sectors like Computer & Math, AI is theoretically capable of handling 94% of tasks[cite: 171]. However, in professional settings, current usage covers only 33%[cite: 174].

Which Jobs are Most Exposed?

The study identifies "Observed Exposure" by looking at where AI is being used for automated, work-related tasks[cite: 13, 123]. Here are the top impacted roles:

Occupation	Observed Exposure	Leading Automated Task
Computer Programmers	74.5%	Writing and maintaining software [cite: 180]
Customer Service Reps	70.1%	Handling complaints and orders [cite: 180]
Data Entry Keyers	67.1%	Entering data into systems [cite: 180]
Medical Record Specialists	66.7%	Coding patient data [cite: 180]
Market Research Analysts	64.8%	Preparing findings and reports [cite: 180]

Conversely, 30% of workers have zero exposure[cite: 189]. These include physically-demanding roles like cooks, motorcycle mechanics, lifeguards, and bartenders[cite: 190].

Early Warning Signs: The Young Worker Effect

While mass unemployment hasn't hit yet, the study found early signs that the "front door" is closing for new graduates[cite: 17, 23]:

• Hiring for workers aged 22-25 has slowed in highly exposed occupations[cite: 17, 271].
• The job-finding rate for young entrants into exposed roles has dropped by approximately 14% since late 2022[cite: 17, 277].
• Younger workers are relatively less likely to be hired into exposed occupations, while job finding rates in less exposed roles remain stable[cite: 275, 276].

The Profile of the "Exposed" Worker

Impacted workers often differ significantly from those in unexposed roles[cite: 223, 224]:

• Higher Paid: Exposed workers earn 47% more on average[cite: 225].
• More Educated: People with graduate degrees are nearly 4x more likely to be in the "most exposed" group than the "unexposed" group[cite: 226].
• Demographics: The exposed group is 16% more likely to be female and significantly more likely to be white or Asian[cite: 224].

Source: Maxim Massenkoff and Peter McCrory, "Labor market impacts of AI: A new measure and early evidence," Anthropic, 2026[cite: 8, 294].