Each Saturday, Matouš Hrdina examines societal trends often overlooked in the daily news cycle.
Debates surrounding the impact of artificial intelligence on the job market have resurfaced with the introduction of new AI tools, prompting familiar questions about potential job losses and economic shifts. Despite recurring cycles of concern, a clear understanding of AI’s true impact remains elusive.
The latest wave of anxiety stems from the launch of new AI tools from Anthropic, which reportedly leverage autonomous “agents” to automate tasks and potentially displace workers across various industries. Following the announcement, shares of software and data analytics companies experienced a significant decline, mirroring past reactions to similar developments.
Dario Amodei, CEO of Anthropic, outlined a potential technological shift in a January essay, while a post by AI entrepreneur Matt Shumer on X garnered widespread attention.
Shumer likened the current AI landscape to February 2020, just before the onset of the COVID-19 pandemic, suggesting that the transformative potential of AI is being underestimated. He acknowledges previous predictions of mass layoffs due to AI, but argues that the current generation of tools represents a fundamentally different level of advancement, potentially ushering in a new economic reality.
But, some critics point out that these cycles of hype often coincide with financially struggling AI companies seeking new funding from investors, and that apocalyptic predictions can also serve as effective public relations tools.
Listen to the audio version of the newsletter read by the author.
AI optimists often cite labor market statistics, but Johannes Kleske and Brian Merchant argue that these arguments are repetitive, often masking underlying managerial decisions, economic cycles, or competitive pressures.
The tendency to attribute job losses to AI, they suggest, can serve to reassure investors and provide a convenient explanation for displaced workers. Kleske also notes a lack of evidence suggesting increased productivity or free time among those utilizing AI in their work, referencing the Jevons paradox, which posits that increased efficiency can lead to greater consumption.
The debate remains inconclusive, with both optimists and pessimists relying on data that can be interpreted in multiple ways. While AI undoubtedly displaces some jobs, its impact is uneven and experiences vary widely across companies.
Where We Excel
The discussion often gets stuck in a loop, focusing on whether AI can replace specific roles, rather than critically examining the purpose and value of those positions within organizations and the broader economy.
Early concerns about AI’s impact on the job market drew comparisons to the historical impact of automation during the Industrial Revolution. While the analogy to the Luddites remains relevant, it’s an imperfect comparison.
The economic consequences of technological automation are now manifesting in creative and skilled professions, such as translators, graphic designers, and, to some extent, programmers. However, the current economic landscape differs significantly from the early 19th century, with a large portion of the workforce employed in roles that don’t resemble the manual labor of a textile factory.
These roles, often described as “bullshit jobs” by anthropologist David Graeber in his 2013 viral essay, are characterized by a lack of clear economic benefit or contribution to society.
Graeber’s work, recently published in Czech translation by Malvern, provides a framework for understanding the current debate surrounding AI and employment. He defines bullshit jobs as those that are not only unnecessary but also recognized as such by both the employee and outside observers, requiring a degree of pretense and self-deception.
He categorizes these jobs into five types: flunkies (who exist to make their superiors sense significant), goons (who engage in deceptive or adversarial activities), duct tapers (who provide temporary fixes to systemic problems), box tickers (who create the illusion of productivity through bureaucratic processes), and taskmasters (who manage and coordinate without adding substantial value).
Interestingly, many bullshit jobs are well-suited for automation by AI, yet their complete elimination seems unlikely. Their primary function is often to create the *appearance* of work, rather than contributing tangible value.
AI tools may even reinforce these positions by enabling employees to simulate productivity. As Will Manidis argues, AI can function as “Tool Shaped Objects,” symbolic tools that create the illusion of work without actually producing meaningful results.
This dynamic highlights a key point: the demand for *feeling* productive often outweighs the need for actual productivity. AI companies are eager to cater to this demand, providing tools that create a sense of accomplishment even if they don’t significantly improve output.
The scenario suggests that the future impact of AI on the job market may not be as dramatic as some predict. Rather than widespread displacement, AI may become integrated into existing “bullshit jobs,” reinforcing a status quo where the appearance of work is prioritized over actual productivity.
