Instead of cutting BS jobs, AI floods us with “workslop”
Even artificial intelligence (AI) may not be able to take over “bullshit jobs.” Instead, it can perform meaningless tasks more efficiently.
Anthropologist David Graeber’s theory of “bullshit jobs” refers to the jobs that even people who perform them think are meaningless.
These are generally office jobs, especially managerial positions, whose importance is often defined by how many people report to these managers rather than the actual work done.
While there isn’t robust data to support the “bullshit job” theory, it resonated among thousands of workers who shared stories of how they spend hours doing meaningless tasks or pretending to be working.
One in five workers in the United States says their jobs don’t make meaningful contributions to the world, whereas 37% of working British adults may be stuck in these BS jobs.
Widespread AI adoption has raised concerns about the technology replacing humans at work. However, “bullshit jobs” that Graeber called the form of spiritual violence may not be the first ones to be eliminated.
The risk of “efficient inefficiency”
According to a 2025 paper, AI adoption may lead to “efficient inefficiency” if the technology is used to perform unproductive, superfluous tasks. In this case, it doesn’t matter if AI can perform these tasks more efficiently.
The authors argue that senior managers who decide how technology is used don’t know how junior managers and workers will respond to their decisions, making it difficult for them to determine superfluous tasks from productive ones.
Where employers tolerate inefficient functions for the sake of status, the use of AI alongside workers may add to inefficiency rather than reduce it. Bullshit jobs in the way Graeber defines them may persist in this case.
David Spencer
Because junior managers don’t know why senior managers are making particular decisions, they are incentivized to increase uncertainty as a defensive strategy. This includes creating new superfluous tasks and not disclosing existing ones.
David Spencer, a professor at Leeds University and co-author of the paper, says there is a risk that AI will create more meaningless work.
“Where employers tolerate inefficient functions for the sake of status, the use of AI alongside workers may add to inefficiency rather than reduce it. Bullshit jobs in the way Graeber defines them may persist in this case,” he tells Cybernews.
Nearly one in five (19%) of employees report frequent use of AI at work, which nearly doubled from 2023. However, real-life data suggests AI may not bring the expected productivity.
According to a 2024 study, using AI copilots resulted in more code written, but it also led to more code churned, meaning it was reverted, removed, or updated within two weeks.
“Workslop” is a new term for low-effort AI-generated content that looks good but lacks substance or “masquerades as productivity,” as Harvard Business Review puts it. Forty percent of US employees reported receiving workslop last month.
Louis Carter, founder of Most Loved Workplace, says automating repetitive work, such as reporting, coordination, and documentation, exposes which roles truly matter.
He says AI can be used for sentiment and thematic analysis to listen better to the workforce and help make quicker decisions on how best to lead and manage for maximum efficiency.
“The danger is when leaders use AI as a command and control mechanism. That’s when it breeds new meaningless work – people managing dashboards with meaningless data instead of relationships,” Carter says.
Shawn DuBravac, CEO and President at Avrio Institute, says AI can reduce “bullshit jobs” by automating coordination and compliance chores. Agentic systems routing tasks end-to-end could help shrink the need for middlemen to pass information across organizational silos.
Workers are increasingly doing repetitive shadow work because of AI. Many are essentially doing the same work twice. They do it manually and then they also do it using AI ‘just to be safe.’
Shawn DuBravac
However, AI could increase “bullshit jobs” by creating environments where humans need to chaperone AI models and agents.
DuBravac says poorly designed systems can create work where humans simply retype AI output without improving decision-making or increasing value.
“Workers are increasingly doing repetitive shadow work because of AI. Many are essentially doing the same work twice. They do it manually and then they also do it using AI ‘just to be safe,’” DuBravac told Cybernews.
Will AI allow us to work less?
One of Graeber’s proposed solutions to “bullshit jobs” is a four-day work week, which has been associated with reduced stress, burnout, fatigue, and work-family conflict.
Of 61 companies that participated in a 2022 trial putting their employees on a four-day work week without cutting pay, 56 are continuing the policy. Besides considerable benefits for employees’ work-life balance, 35% of these companies also reported revenue growth.
A 2024 survey of senior leadership found that 93% of organizations where AI plays a central role in operations are either considering or have already implemented a four-day working week.
However, Spencer says widespread AI adoption may not lead to a reduction in working hours because AI maximizes returns for employers.
The key to making AI work for people – and for more meaningful work – is to ensure that it is democratically owned and directed. Without greater democracy over its development and use, we risk a future where AI benefits a few and leaves the majority without the means, in work or beyond it, to live well.
David Spencer
For example, using AI for monitoring may lower costs for employers but at the expense of more intensive work for workers.
For now, AI isn’t driving revenue growth as many may have expected. A recent MIT Media Lab report suggests that 95% of organizations are getting zero return on generative AI investments.
Spencer says automation may reduce headcount rather than improve the quality of work and reduce working hours. Nearly half (43%) of business leaders expect to replace junior roles or entry-level jobs with AI, including by cutting jobs.
He tells Cybernews, “The key to making AI work for people – and for more meaningful work – is to ensure that it is democratically owned and directed. Without greater democracy over its development and use, we risk a future where AI benefits a few and leaves the majority without the means, in work or beyond it, to live well.”
Unlock more exclusive Cybernews content on YouTube.
