AI Misuse Creates Unnecessary Work and Damages Trust in Teams

A recent study highlights a significant issue in workplaces where artificial intelligence (AI) is employed: many employees are relying on AI-generated content without verifying its accuracy. Research indicates that approximately two-thirds, or 66 percent, of workers use AI outputs without evaluation, leading to increased workloads and potential reputational damage. A notable example occurred this week when Deloitte Australia issued an apology after a report valued at A$440,000 (US$284,000) for the federal government was found to contain multiple AI-generated errors.

The Emergence of ‘Workslop’ in the Workplace

This phenomenon has given rise to the term “workslop,” which was popularised in a recent article in the Harvard Business Review. It describes AI-generated content that may appear polished but lacks the necessary substance to effectively advance tasks. A survey conducted by BetterUp Labs and the Stanford Social Media Lab revealed that 40 percent of U.S. workers reported receiving workslop from colleagues in the past month. Each instance, on average, took nearly two hours to rectify, leading to an estimated annual productivity loss of around US$9 million for a company with 10,000 employees. Workers receiving such content expressed feelings of annoyance and confusion, often perceiving the sender as less reliable or trustworthy.

Understanding the Human Impact of AI Overreliance

Research conducted by the Melbourne Business School found similar patterns of complacency regarding AI use. A survey involving 32,352 workers across 47 countries revealed that while some employees experienced improvements in efficiency or innovation from AI, over a quarter noted an increase in workload and pressure from mundane tasks. Alarmingly, half of the respondents admitted to using AI instead of collaborating with colleagues, raising concerns about the future of teamwork.

The study also indicated that many employees are secretive about their AI usage; 61 percent of respondents avoided disclosing their use of AI, while 55 percent presented AI-generated material as their own work. This lack of transparency complicates efforts to identify and correct errors produced by AI.

Strategies to Mitigate Workslop

To address the issue of workslop and optimise the benefits of AI, specific steps can be taken by both employees and employers. Employees can adopt a few straightforward strategies to improve the quality of AI-generated work.

Employers, on the other hand, must invest in governance, AI literacy, and skills for human-AI collaboration. Establishing clear guidelines for effective AI use is essential. This includes formulating an AI strategy that identifies high-value applications, defining roles and responsibilities, and monitoring outcomes. Proper governance can reduce risks and minimise the rework associated with workslop.

As workslop stems from how individuals use AI rather than the technology itself, effective governance needs to shape daily behaviours. Closing the AI literacy gap is crucial; research shows that better training correlates with more critical engagement with AI and fewer errors. Nonetheless, less than half of employees report having received any form of training or policy guidance regarding AI use.

To diminish workslop, employees need to learn how to use AI selectively, responsibly, and collaboratively. They should also be taught methods for verifying AI outputs before sharing them with others, thus enhancing overall productivity and trust within teams.

The insights shared in this article reflect the collaborative research efforts of Steven Lockey, a Postdoctoral Research Fellow at Melbourne Business School, and Nicole Gillespie, Chair in Trust and Professor of Management at The University of Melbourne. Their findings underscore the importance of addressing AI misuse in the workplace to foster a more productive and trustworthy environment.