AI is creating ‘workslop’ and hindering workplace productivity

Companies are mandating the use of generative AI (Gen AI), and employees are adopting it to increase efficiency.

However, a new study shows that this is leading to a large volume of low-quality output, termed “workslop,” which creates a greater workload for the end recipient and increased costs for businesses.

Researchers from the Harvard Business Review, in collaboration with the Stanford Social Media Lab, coined the term “workslop” to describe this new problem.

Workslop takes the concept of “AI slop” found on social platforms and applies it to items like poor-quality reports and slides produced by individuals who exert minimal effort before feeding prompts to their AI tools.

The research found that this disproportionately affects the technology sector. However, 40% of survey respondents said they had encountered workslop in the past month.

Those who reported receiving workslop said that an average of 15.4% of the work they received fit the definition of slop.

While workslop is mostly a peer-to-peer problem (40%), 18% of it goes from direct reports to their managers.

HBR’s survey of 1,150 full-time employees across various sectors in the US reveals the prevalence of workslop.

Workslop also carries an emotional and social cost. When asked how it felt to receive workslop, 53% of employees reported feeling annoyed, 38% felt confused, and 22% felt offended.

In addition to having to review workslop for errors and missing information, recipients also experience the stress of having to diplomatically discuss the issue with their colleagues and bosses.

The study estimates that for a company with 10,000 employees, these work errors lead to an annual productivity loss of $9 million.

The HBR study’s findings appear consistent with a July study from the MIT Media Lab, which revealed that despite a $30 to $40 billion investment in Gen AI tools, 95% of organizations have failed to see any return from the technology.