A new study warns that intensive use of AI tools is causing a distinct form of mental exhaustion among workers, separate from traditional burnout. Researchers at Boston Consulting Group report findings from a survey of 1,488 full-time U.S.-based workers across industries, roles, and seniority levels.
The study identifies a phenomenon the researchers call “AI brain fry,” defined as mental fatigue from excessive use or oversight of AI tools beyond a person’s cognitive capacity. Fourteen percent of workers using AI on the job reported experiencing it. Symptoms include difficulty focusing, slower decision-making, headaches, and a general sense of mental fog.
The research finds that the most mentally taxing aspect of AI use is not the volume of tasks but the oversight required. Workers who reported high levels of AI oversight showed 14% more mental effort, 12% more mental fatigue, and 19% greater information overload compared to those with low oversight demands. When AI tools also increased overall workload, cognitive strain rose further.
The number of tools used simultaneously matters too. Productivity increased as workers moved from one to two tools, and again with a third. After three tools running at once, however, productivity scores dropped. The researchers describe this as a multitasking trap that workers fall into despite its well-documented inefficiencies.
AI brain fry carries real business consequences. Workers experiencing it report 33% more decision fatigue than those who do not. They also make mistakes more often, scoring 11% higher on minor errors and 39% higher on major errors. Among workers who did not report AI brain fry, 25% showed intent to leave their job. That figure rose to 34% among those who did, a 39% increase in turnover risk.
Not all AI use drives mental fatigue. When workers use AI to eliminate repetitive or routine tasks, burnout scores fall by 15%. Those workers also reported higher engagement, more positive feelings about AI, and stronger social connections with colleagues.
The study highlights that individual choices alone do not determine outcomes. Manager and team behaviour plays a significant role. Workers whose managers actively answer AI-related questions show 15% lower mental fatigue. When managers leave employees to figure out AI independently, fatigue scores rise by 5%. Organized team-level integration of AI reduces mental strain, while team pressure to use AI increases it.
At the organizational level, unclear communication about AI’s role in workload correlates with higher fatigue. Workers who feel their employer values work-life balance report 28% lower mental fatigue scores.
The researchers offer several recommendations:
- Organizations should limit the number of AI agents any one person oversees.
- They should set explicit expectations about workload rather than implying intensification through productivity messaging.
- Metrics should focus on outcomes rather than activity.
- Workers need training in skills such as problem framing and strategic prioritization to avoid unnecessary AI-driven work loops.
The researchers conclude that the difference between beneficial and harmful AI use lies not in how much AI a person uses but in how individuals, teams, and organizations shape that use.
