Advances in natural language processing and generative artificial intelligence have sparked new fear among white-collar professionals long wary of the automation of so-called knowledge work. The launch of ChatGPT late last year prompted a fresh wave of articles speculating on how soon copywriters, lawyers, programmers, and others might find their professions redundant.
The real question, however, is why more such work wasn’t automated long ago. Peter Drucker, the business guru who coined the term “knowledge work” in 1959, characterized it as “ever-changing, dynamic, and autonomous,” but in reality, much of what falls into that category is nothing of the sort. Many “knowledge workers”—from lawyers to accountants, marketers to engineers, assistants to supervisors, administrators, writers, editors, content producers, software developers, and all of the tedious email jobs in between—perform repetitive tasks that could be entrusted to technologies far less sophisticated than neural networks. Trend pieces on professionals collecting a paycheck for a job they have secretly automated prove the point.
In his 2018 book Bullshit Jobs, the late anthropologist David Graeber examined the phenomenon of jobs—mostly white-collar—so devoid of utility and purpose that even those working them can’t justify their existence. The proliferation of such jobs, he argued, contradicts the common belief that our economy is guided by market logic. It’s supposed to be stagnant state-managed economies that generate pointless make-work jobs, not lean capitalist ones guided by the imperatives of productivity and efficiency. How could this have happened?