The Burden of Botsitting: A New Layer of Work
Imagine hiring an assistant to help manage your email. Initially, your time investment skyrockets as you spend weeks explaining filtering preferences and correcting mistakes. By the end of the month, instead of relief, you find yourself managing not only your email but also supervising your assistant’s work. This scenario encapsulates the phenomenon of “botsitting,” a term that describes the additional workload created by relying on AI tools without experiencing actual time savings.
The Reality of AI at Work
According to the Work AI Index from the Glean Institute, employees are averaging 6.4 hours weekly just to make AI functional. That’s nearly a full workday lost, drowning amidst the promise of automation. Despite 87% of employees reporting that they utilize AI, only 13% of companies see real productivity gains. The discrepancy between individual perceptions and organizational outcomes illustrates a significant gap; rather than disappearing, time is merely redirected to a new layer of responsibilities that often go unnoticed.
Understanding Botsitting
The task of botsitting involves acting as an “AI babysitter,” managing the context, reviewing the AI’s outputs, and correcting its mistakes. As Rebecca Hinds from the Work AI Institute notes, this babysitting is “tedious and exhausting work,” stripped of acknowledgment or reward. Thus, saved AI time morphs into an obligation that employees must fulfill at later hours.
The Challenge of Too Many Tools
The complications of botsitting are exacerbated by the overabundance of AI tools. Around 77% of employees report using multiple AI applications weekly, with a significant portion toggling between four or more tools. Each transition brings a cognitive cost, commonly referred to as the “toggle tax,” where workers waste time repeating prompts and instructions. According to studies by the Harvard Business Review and McKinsey, employees lose nearly two hours daily searching for information across various systems. Far from easing workflow, AI has added to the disarray.
Botshitting: The Consequence of Overload
As employees grapple with correcting AI errors under tight deadlines, a looming threat emerges: botshitting, or delivering unverified AI-generated work. A staggering 69% of participants admitted to submitting such unreviewed work at least occasionally. The implications are far-reaching; the burden of correcting unchecked output is then shifted to the next link in the chain, perpetuating inefficiencies and costs across the board.
Adding More AI Is Not the Answer
Bob Sutton, a Stanford professor, argues that management often misjudges the solution to AI-related friction. Instead of resolving misuse through additional AI tools, organizations should focus on fostering “human infrastructure.” Data shows that companies excel not by overusing AI but by ensuring that useful information flows effectively. Where this is prioritized, employees report feeling 64% less exhausted and are significantly less likely to submit work that hasn’t been reviewed.
As the landscape of work continues to evolve, it’s essential to reassess the role of AI not merely as a time-saving tool, but as a complex layer of responsibilities requiring careful management. In navigating this paradox of botsitting, companies must balance technology with human insight to truly enhance productivity and employee well-being.