Research Warning: Overuse of Artificial Intelligence May Trigger "AI Brain Fatigue," Leading to a Significant Increase in Employee Error Rates and Turnover

March 9 News, a recent study shows that widespread use of artificial intelligence tools in the workplace has not fully alleviated employee stress; in some cases, it has led to new cognitive burdens. Researchers from Boston Consulting Group and the University of California published a report in the Harvard Business Review stating that over-reliance on AI may cause employees to experience so-called “AI brain fatigue.”

The study surveyed nearly 1,500 full-time employees in the United States, with about 14% reporting noticeable mental fatigue from frequent use or supervision of AI tools. This phenomenon, termed “AI brain fatigue” by the researchers, mainly manifests as a state similar to “mental hangover,” including feelings of mental fog, decreased attention, slowed thinking, and difficulty making decisions.

The research indicates that as companies deploy more AI tools and multi-agent systems, employees need to switch between different platforms frequently, increasing cognitive load. Originally promoted as efficiency boosters, AI assistants sometimes complicate task management in practice. Some companies even include AI usage frequency as a performance metric, further encouraging employees to increase their usage.

For example, Brian Armstrong, CEO of the US-based compliant CEX, publicly stated that the company is encouraging engineering teams to heavily use AI and plans to have AI generate about half of the software code. Such corporate strategies accelerate AI’s integration into workflows.

The results show that employees experiencing “AI brain fatigue” are about 33% more likely to suffer decision fatigue and approximately 40% more likely to consider leaving their jobs. Additionally, the likelihood of making significant errors—defined as mistakes that could impact safety, business outcomes, or key decisions—is nearly 40% higher. The report notes that such major errors could potentially cost large enterprises millions of dollars annually.

However, the researchers also pointed out that proper use of AI can still bring positive effects. If AI is mainly used for repetitive and routine tasks, such as data organization or process automation, employee burnout levels can be reduced by an average of about 15%.

The research team recommends that companies clearly define the boundaries of AI application when promoting these tools and focus on improving work outcomes rather than simply increasing usage frequency. By designing AI workflows thoughtfully, organizations can leverage technological advantages while avoiding additional mental stress caused by overuse of AI.