AI tools do not reduce workloads — they tend to increase them. That is the conclusion of a new study published by Harvard Business Review (HBR).
Over an eight-month period, researchers observed how employees at a U.S.-based technology company with roughly 200 staff members adapted their work habits as AI became integrated into routine processes.
The findings showed that as artificial intelligence was embedded into daily workflows, employees extended their working hours, operated at a faster pace, often without direct managerial pressure, and took on a broader range of responsibilities:
-
Product managers and designers began writing code.
-
Researchers handled engineering-related tasks.
-
Employees across the organization attempted work they would previously have outsourced, postponed, or avoided entirely.
Expanded responsibilities and side effects
The expansion of employee scope had unintended consequences. Developers, for example, spent more time reviewing, correcting, and refining outputs produced by colleagues using AI. They increasingly acted as mentors to coworkers engaging in so-called “vibe coding.”
Blurred boundaries
Because AI significantly lowers the barrier to starting a task, employees began initiating work during moments that were previously considered breaks. Many launched AI tools during lunch, meetings, or while waiting for files to load.
Some even submitted a “quick prompt” before leaving the office so the LLM could continue working in their absence.
These behaviors were not initially perceived as additional workload. Over time, however, the workday became less structured and more continuous. The conversational nature of prompting further softened the experience — typing instructions felt more like chatting than performing formal job duties. This made it easier to shift work into evenings or early mornings without consciously intending to extend the day.
“The boundary between work and non-work has not disappeared, but it has become easier to cross,” HBR noted.
Multitasking intensifies
AI introduced a new rhythm in which employees managed multiple active streams simultaneously: manually writing code while AI generated alternatives, running several agents in parallel, or revisiting long-postponed tasks.
Workers did this because they felt they had a “partner” helping them manage the load. However, this dynamic led to constant attention switching and a persistent sense of busyness.
Driven by enthusiasm
The company did not mandate AI adoption. Employees embraced it voluntarily because the technology allowed them to “do more.”
In the short term, HBR notes, this trend appears beneficial for management. But enthusiasm for experimentation may fade, and employees may eventually realize their workload has significantly increased.
“This expansion of activity can lead to cognitive fatigue, burnout, and weakened decision-making capacity. The initial productivity surge may give way to declining work quality, higher turnover, and other issues,” the study warns.
Researchers emphasized that the situation places leaders in a difficult position. Expecting employees to self-regulate their workload is not an effective strategy. Companies should instead establish clear norms and standards for AI usage.
“Without such practices, the natural tendency of AI-assisted work is not reduction, but intensification — with consequences for burnout, decision quality, and long-term sustainability,” the report concludes.
Recommendations
HBR suggests implementing several measures:
-
Intentional pauses to prevent workload accumulation.
-
Sequential work practices — delaying non-urgent notifications and updates to protect focus and reduce interruptions.
-
Structured in-person interaction — allocating time and space for human connection, including short meetings, collaborative reflection, or organized dialogue.
ES 
EN