Ken, a copywriter at a prominent cybersecurity company in Miami, once found satisfaction in his role. However, the emergence of what’s termed “workslop” has significantly altered his work experience.
Workslop refers to the unintended fallout from the rapid adoption of artificial intelligence in the workplace. It manifests when employees utilize AI to produce outputs that appear refined on the surface but are riddled with inaccuracies and errors, necessitating extensive revisions or even complete reworking once shared with colleagues.
For Ken, the challenges began after the CEO of his firm implemented layoffs and mandated that the remaining staff utilize AI chatbots, claiming this would enhance productivity. Although generating initial drafts became easier, Ken and his team discovered they spent more time revising and reconciling discrepancies among their AI-generated texts than they would have if they had created the content without AI assistance.
“The quality of work dropped dramatically, the time taken to produce content increased, and most critically, employee morale suffered,” Ken explained, opting for anonymity due to concerns about job security. “The situation worsened significantly following the implementation of AI.” He noted that executives often deflected criticism onto employees when productivity declined.
Ken’s situation highlights a growing rift between employees and management regarding AI usage. A recent survey involving 5,000 white-collar workers in the U.S. revealed that 40% of non-managerial employees felt AI did not save them any time, in stark contrast to 92% of executives who believed it enhanced productivity.
The root of the workslop issue is multifaceted and cannot be attributed solely to workers cutting corners. The core problem is tied to decisions made at the executive level.
Companies have invested billions in generative AI technologies, and some, including Block, Amazon, and Target, have simultaneously reduced their workforce, citing AI’s potential for increasing productivity. Remaining employees often feel pressured to produce more with AI tools, frequently without adequate guidance or training. This disconnect highlights the contrasting perspectives of enthusiastic executives and employees who find their tasks more burdensome due to AI.
“Employees are being instructed to use AI tools but often lack proper support or clear instructions,” remarked Jeff Hancock, a co-author of the study that introduced the term “workslop” and a researcher at Stanford University. He acknowledged that while generative AI has the potential to enhance productivity, its current implementation often yields the opposite result.
Hancock’s study, which is not yet peer-reviewed, surveyed 1,150 desk workers from the larger pool of 5,000. The findings indicated that 40% of participants had experienced workslop within a month, spending an average of 3.4 hours monthly addressing it. This inefficiency translates to an estimated $8.1 million in lost productivity for organizations with around 10,000 employees.
Kelly Cashin, a freelance product designer, shared her frequent encounters with workslop. “It has become commonplace to directly copy and paste responses generated by AI into communications,” she noted. When faced with confusing messages from colleagues, responses often reflect uncertainty about AI’s intent. “While it can be frustrating, I understand the pressure to be productive, especially in an uncertain job market,” Cashin added.
Philip Barrison, an MD-PhD student at the University of Michigan, observed similar issues with medical staff using AI to draft email responses to patient inquiries, intended to save time for clinicians. However, he found that this approach often resulted in increased editing work and concerns over data security, leading many to disregard the AI tools after the novelty wore off.
Employers are promoting generative AI to cut labor costs following significant investments in this technology, according to Aiha Nguyen, head of the Labor Futures program at the Data & Society research institute. However, these investments have not yet yielded the anticipated returns. A widely referenced MIT report indicated that 95% of companies are not experiencing a return on their AI investments. Other reports, from SAP and Deloitte, suggest a greater number of businesses are seeing returns, yet this remains a minority. Companies are optimistic that better results will emerge within two to four years, a timeline considered slow for technological investments.
“The challenge lies in the perception of generative AI as a universal tool, while reality shows it does not function that way. The unclear guidelines surrounding AI use may contribute to the creation of workslop,” Nguyen explained.
AI has become a contentious issue as unionized workers negotiate new contracts, according to Dan Reynolds, a research economist with the Communications Workers of America. Unions are advocating for clearer guidelines on technology use and more employee involvement in its implementation.
“Companies are candid about their intentions to use AI to streamline processes, prompting necessary scrutiny of what these tools can actually accomplish and the underlying power dynamics,” noted Sarah Fox, director of the Tech Solidarity Lab at Carnegie Mellon University.
