Boston Consulting Group: Excessive AI Drains Minds

An Alert Issued by the Boston Consulting Group

Researchers from the Boston Consulting Group have recently highlighted a concerning phenomenon affecting workers who simultaneously use multiple artificial intelligence (AI) tools. They have observed that some employees suffer from a condition they call "AI-related brain fatigue."

Symptoms and Impact on Productivity

This brain fatigue manifests as mental fog and other signs of mental exhaustion, resulting from excessive mental strain. Although AI tools are designed to enhance productivity, their overuse can have the opposite effect. A recent study published in the Harvard Business Review surveyed 1,488 full-time workers in the United States from large companies across various sectors to explore this phenomenon.

The study revealed that a significant number of workers experience symptoms such as mental fog, headaches, and a slowdown in decision-making. These symptoms are grouped under the term "AI-related brain fatigue."

Testimonials and Researchers' Observations

Matthew Kropp, managing director at Boston Consulting Group and one of the lead authors of the study, compared this situation to a "canary in the coal mine." He explains that engineers, often early adopters of new technologies, are already feeling this effect. As more people adopt these technologies, this phenomenon could intensify.

The Limits of Multitasking with AI

The study emphasizes that while AI tools can indeed boost productivity, this is only true up to a certain point. A senior engineer shared his personal experience, explaining that he used multiple tools for different tasks, but this ultimately cluttered his mind rather than helping him move faster.

The results show that employees using one or two AI tools simultaneously notice a significant increase in their productivity. However, this increase becomes less significant when they switch to three tools, and it declines beyond that number, illustrating the limits of multitasking.

Managing AI Agents: A Cognitive Challenge

Kropp insists that this is not about discouraging the use of multiple AI agents but rather about being aware of the cognitive impact and managing it. He distinguishes between general burnout and AI-specific fatigue, which arises from overseeing agents that require a high cognitive load.

Consequences for Companies and Recommendations

The stress associated with managing multiple AI agents can reach a breaking point for workers. Therefore, companies need to be mindful in developing their AI policies, particularly by allowing breaks for employees.

Kropp suggests that increased productivity should not come at the expense of employees' mental health. He proposes that even if an employee is 50 times more productive, it might be preferable for them to be 20 times more productive while preserving their mental well-being.

Limiting AI Tool Usage Time

Steve Yegge, a veteran software engineer, recently recommended limiting the use of AI-assisted coding tools to three hours per day for engineers.

The Adoption of AI and Its Variable Effects

The emergence of AI-related brain fatigue does not mean that companies should avoid AI. The study shows that when AI replaces routine tasks, the rate of burnout decreases, even though mental fatigue persists.

Kropp explains that learning and using these tools is rewarding, but it is crucial to manage their negative effects.

Variability of Fatigue Across Sectors

The study found that 14% of surveyed workers experience this AI-related brain fatigue, with significant variations across sectors. Employees in marketing (25.9%), human resources (19.3%), operations (17.9%), and software engineering (17.8%) are more affected than those in legal and compliance sectors, where the rate is around 6%.

Kropp attributes these differences to the rates of AI adoption in each sector. He notes that the capabilities of AI agents have evolved significantly over the past nine months, and other sectors are expected to soon experience changes similar to those observed in software engineering.

"In corporate organizations, less than 5% of workers currently use multiple AI agents," Kropp points out, emphasizing that we are still at the beginning of this transformation.