How Can AI Create Cognitive Friction?

One of the most compelling promises of generative AI is that it will remove tedious work from overloaded human workers. AI can draft the email, summarize the document, organize the information and automate the repetitive task, leaving people with more time for the work that requires those uniquely human skills like judgment, creativity, strategy…the list goes on.

This sounds like an obvious win. But there is an important question for employers: what happens when the work we removed was also giving our brains a chance to recover?

This is where cognitive offloading becomes particularly interesting. We often assume that removing tasks will automatically reduce mental strain, yet not all easy work is wasted work. Routine, familiar tasks can provide periods of lower-intensity cognitive activity between more demanding stretches of concentration. If AI eliminates all of these moments, an employee’s day can become an uninterrupted sequence of high-value, high-attention tasks that ultimately reduce productivity and increase worker fatigue.

The result may therefore be a paradox…AI can make every hour more productive and more mentally demanding, which can then result in lower quality, less effective work.

There is already evidence that AI is changing how workers experience their jobs. The PwC Global Workforce Hopes & Fears Survey reports that around a third of workers feel overwhelmed at least once a week, while only 14% of respondents say they use generative AI daily. At the same time, daily AI users report significant benefits in productivity and job security. So the lesson isn’t that AI is making people miserable, it is that productivity gains do not guarantee cognitive wellbeing.

Then there is the added layer of continuous learning. Generative AI is evolving rapidly, and employees are being asked not simply to use a new tool but to continually learn new ways of working. Which model should I use? What should I automate? How do I prompt effectively? Can I trust this output? What has changed since last month?

The European Union’s AI Act recognizes the importance of this learning curve. Article 4, which began applying in February 2025, requires organizations deploying AI systems to take measures to develop AI literacy among their staff. That is an important step, but policy and training alone cannot solve cognitive strain. People also need space to absorb change.

And then there is the emergence of FOBO: the Fear of Being Obsolete. As AI becomes capable of performing more tasks once considered valuable human expertise, employees may understandably wonder whether their skills will remain relevant. That anxiety can turn learning into a permanent defensive exercise; I need to keep up, I need to learn the next tool, I need to prove my value.

For employers, therefore, the question shouldn’t simply be, “How much work can AI remove?”. It should be, “What kind of work are we replacing, and what are we filling the space with?”.

If every saved minute becomes another meeting, another analysis, another deadline or another expectation, AI hasn’t reduced cognitive strain. It has simply increased the intensity of the working day as well as the likelihood that workers will experience cognitive overload from too many immediate and complex demands on the brain.

AI has extraordinary potential to make work better. But a sustainable AI strategy must recognize something very human: our brains need recovery as well as productivity. Sometimes the seemingly unimportant task isn’t getting in the way of better work. It might be helping us get ready for it.

If you would like to discuss how we can help ensure that AI is embedded into your company with the right balance for your employees’ wellbeing, please get in touch with me at amanda@orgshakers.com or through our website here.