AI promised productivity gains, workers report ‘workslop’
NEW YORK, UNITED STATES — A growing share of American white-collar workers say generative AI is not making them faster — it is creating a new category of low-quality output their colleagues then have to fix, a phenomenon researchers are calling “workslop” that is costing large employers millions of dollars in lost productivity.
A new Stanford and BetterUp study, surveying 1,150 United States desk workers, found that 40% encountered workslop within a single month and spent an average of 3.4 hours fixing it — translating to an estimated $8.1 million in lost productivity for a 10,000-person organization.
The widening gap between executives and workers
The data exposes a sharp divide inside U.S. companies. A separate survey of 5,000 white-collar workers found that 92% of high-level executives say AI makes them more productive, while 40% of non-managers say it saves them no time at all.
Workers describe being told to use AI tools without training, then blamed when output quality falls.
Companies including Block, Amazon, Dow, UPS, Pinterest and Target have cut jobs while citing AI’s productivity potential, leaving remaining staff under pressure to produce more with tools they were never trained to use. The result, according to the study, is more rework — not less.
“People are being told to use AI, often without direction or support,” said Jeff Hancock, a Stanford researcher and BetterUp scientific adviser who co-authored the study that coined the term “workslop.”
Why the AI investment math isn’t adding up
The productivity gap is colliding with a tougher financial picture.
A widely cited MIT report found that 95% of firms are not seeing returns on their generative AI investments, and follow-up assessments from SAP and Deloitte show only a minority of companies generating measurable gains.
Deloitte projects that better returns may take two to four years to materialize — a long horizon for technology spending.
Labor researchers say the disconnect points to deeper issues with how AI is being deployed in workplaces, including unclear use cases and a tendency to position AI as a general-purpose tool.
Unions, meanwhile, are pushing for clearer mandates and more worker input into how the technology is rolled out.
“The problem is, generative AI is often being presented as a general-use tool that can do anything, but the reality doesn’t work that way. So what could be creating part of the workslop is [AI’s] unclear mandate or use case,” said Aiha Nguyen, who leads the Labor Futures program at the Data & Society Research Institute.
For the outsourcing industry, the workslop trend opens a clear opportunity. As U.S. companies struggle to capture promised AI gains, business process outsourcing (BPO) providers offering trained human-in-the-loop teams — for quality assurance, prompt engineering, content review and AI output validation — are positioned to absorb the rework load that internal staff are buckling under, turning a productivity problem into a defined service line.
