Study: AI Increases Mental Load and Fatigue at Work

Constant supervision of artificial intelligence can increase worker mental effort by 14% and information overload by 19%, according to research published by the Harvard Business Review. While AI aims to simplify tasks, the cognitive demand of monitoring machine-generated outputs often creates new forms of professional fatigue and decision-making errors.

Why does AI supervision cause mental fatigue?

The mental strain stems from a shift in how professionals spend their time. Rather than performing core tasks, many employees now act as auditors for machine output. According to a study by the Boston Consulting Group (BCG) involving 1,488 full-time workers in the United States, those required to constantly monitor AI tools experienced significantly higher cognitive loads.

The research highlights three specific negative outcomes for high-supervision roles:

• 14% increase in total mental effort.
• 12% increase in mental fatigue.
• 19% increase in information overload.

The Harvard Business Review notes that the problem isn’t the technology itself. Instead, it is the “supervisory burden.” Workers must constantly decide whether to trust a recommendation, cross-reference facts, and switch between different digital platforms. This constant state of verification prevents the deep focus required for complex problem-solving.

How can businesses implement AI without increasing burnout?

The impact of AI on the workplace depends entirely on how the tool is deployed. The BCG study found a clear distinction between “supervision-heavy” AI use and “automation-heavy” AI use. When AI is used to handle repetitive, low-level tasks without requiring constant human oversight, the results are positive.

In cases where AI functioned as a background automation tool, researchers observed a reduction in burnout and an overall improvement in the employee experience. The key is to use AI to eliminate drudgery rather than to create a new layer of management tasks for the human worker.

What will the future of AI-human collaboration look like?

As companies integrate more advanced models, the workforce will likely see a shift toward specialized oversight roles. Because high supervision leads to more errors and a higher intention to leave jobs, companies cannot simply dump unverified AI outputs onto employees. Future trends suggest a move toward “Human-in-the-loop” (HITL) systems designed specifically to reduce cognitive friction.

We can expect to see three major shifts in workplace design:

1. Cognitive Load Management

Companies will likely begin measuring “cognitive load” as a key performance indicator (KPI). If an AI tool increases the mental effort of a team by 14%, it may be viewed as a net loss in productivity despite its speed.

2. Higher Standards for AI Accuracy

To reduce the increase in fatigue, the next generation of enterprise AI must focus on “reliability-first” architecture. If a tool is more accurate, the human supervisor spends significantly less mental energy on verification.

3. New Skillsets for Oversight

The role of the professional is changing from a “doer” to a “director.” This requires a different type of training—not just how to use AI, but how to critically evaluate it without becoming overwhelmed by the sheer volume of generated information.

For more insights on how technology is reshaping the modern workplace, explore our latest reports on [Digital Transformation Trends].

Do you feel more or less productive when using AI tools? Have you noticed an increase in mental fatigue? Share your experience in the comments below.