hough many workers worry that AI is going to take their jobs, evidence suggests that it’s actually giving AI adopters more work, not less.

In an eight-month study of approximately 200 workers at a US-based tech company, Harvard University researchers discovered that AI tools consistently intensified work, rather than reducing the load. The researchers found that AI tools allowed workers to complete tasks faster, enabling them to take on a broader scope of tasks, thereby extending their work hours.

Though the company being studied offered enterprise subscriptions to AI tools for their employees, the researchers noted that these employees were not mandated to use AI. Rather, the workers did so of their own accord.

The problem, however, is that once the excitement over these shiny new AI tools wore off, workers found that their workload had increased without them noticing. The researchers identified three main ways that these workloads intensified:

• AI made tasks that were once out of reach feel achievable to new audiences. For example, coding and engineering tasks are now within reach for non-technical employees.
• Reduced friction in starting and completing tasks also blurred the boundaries between work and non-work.
• Finally, these tools allowed for easier multitasking, with the tech being seen as a “partner” that could handle more tasks in the background. The consequence of that, however, was also an increased taskload.

Harvard’s study joins a litany of conflicting research detailing how AI will impact the way we work. While some say that AI can already automate thousands of hours of work and make certain jobs obsolete, others argue that AI will create new jobs entirely. This study lands somewhere in the middle: Creating new work in the jobs that we already have, while quietly piling on more right under our noses.