Why AI’s gains could come with a burnout cost

“Sometimes, I wonder why I bother going to work at all,” an executive in a class I was teaching said during a discussion about artificial intelligence. He was not burned out in the conventional sense; he was struggling to understand where he fit in a workplace where machines were taking on more of the tasks that once defined his job. 

The executive’s anxiety points to one risk of AI adoption: not just heavier workloads, but also uncertainty about the value of human contribution itself. 

That stands in stark contrast to what famed economist John Maynard Keynes posited in 1930 – that the pace of technology development would largely free people from the burden of work within a century. “Three hours a day is quite enough,” he wrote in his famous essay, Economic Possibilities For Our Grandchildren. 

That century is nearly up, and the prediction has not come to pass. Technology, especially AI, has yet to deliver the liberation from work that many anticipated. In some workplaces, it appears to be enabling more work to be packed into every hour. 

Yet Keynes’ vision continues to echo through contemporary discussions about AI. Microsoft co-founder Bill Gates suggested in 2023 that AI could eventually reduce the working week, while Tesla chief executive Elon Musk has gone further, arguing that work may one day become optional. 

Emerging evidence suggests a more complicated reality. A recent study of 136,000 US workers found that those working jobs most exposed to AI were putting in longer hours and reporting lower levels of work-life balance. The study found that workers in AI-exposed jobs logged an average of 3.4 additional hours a week, while time spent on leisure activities declined. The study does not prove that AI is causing burnout. But it does suggest that the way many organisations are adopting the technology may be creating unintended pressures for employees. 

For many workers, AI indeed appears to be increasing mental workload and job complexity before delivering any meaningful time savings. 

And that is turning burnout from a well-being concern into a business risk. An eight-month study of 200 employees at a US technology company found that AI often led workers to take on more tasks and work across more hours of the day. Many felt more productive, but not less busy. The researchers warned that the resulting “workload creep” could contribute to fatigue and burnout. 

Compressed intensity and more multitasking 

AI allows many tasks to be completed more quickly. But in many organisations, additional capacity is not translating into more free time. Instead, it is being filled with new demands. Translators increasingly edit what AI spurts out rather than translating from scratch, while software developers now spend more time reviewing machine-written code. The work has not disappeared; often, it has shifted from creation to supervision. 

And the result is a new form of workplace pressure: compressed intensity. In some roles, AI can fragment work into cycles of prompting, checking and revising generated output. It can also encourage more frequent task-switching. Research has long suggested that sustained concentration and immersion – often described as “flow” – are important sources of motivation, engagement and energy. 

AI tools are also designed to be highly engaging and conversational. And as more work is mediated through machines, employees may find it harder to demonstrate the distinct value of their own contribution. A lawyer might spend 30 minutes spotting the one clause in an AI-generated contract that could create a legal liability. The client sees the finished article, not the judgment that prevented the legal mishap. Output increases, but recognition does not always follow.

AI and burnout

AI does not create a new form of burnout so much as amplify the existing drivers of it. First, it can intensify perceptions of unlimited workload. As employees become more productive, organisations aren’t allowing the “foot off the gas”. On the contrary, many employers respond by raising expectations rather than reducing pressure. The result is that work expands to fill newly available capacity. 

Second, AI can complicate employees’ sense of value and contribution. As machines perform more tasks, some workers begin to question what remains uniquely human about their role. That can undermine the sense of recognition and purpose that sustains motivation. 

Third, work increasingly involves interactions with systems rather than people. If not managed carefully, that may weaken connections with colleagues and managers and contribute to feelings of isolation. 

At the same time, I am hearing a shift among senior leaders away from post-pandemic well-being priorities and towards a renewed emphasis on performance and delivery. Flexible arrangements that became common during and after Covid-19 are being reassessed, while expectations around speed and productivity are rising. Companies including Amazon and JPMorgan Chase, for instance, have pushed employees back into the office more regularly. 

This shift comes at a time when employee well-being remains under strain. Fresh evidence from Singapore points to a workforce under pressure. A survey of 1,000 workers by Telus Health found that 41 per cent reported lower productivity linked to mental health, while more than half were either thinking of leaving their jobs, or were uncertain about staying put. 

Interestingly, employees at organisations that discouraged AI use reported poorer mental health, scoring more than 11 points below those whose companies promoted its adoption. 

The finding may not suggest that limiting AI harms mental health. Rather, it may indicate that firms with more restrictive, prevention-oriented cultures tend to have lower levels of employee well-being, regardless of their approach to AI. 

Burnout was already a significant workplace issue before widespread AI adoption. Now, the concern is that AI could intensify existing pressures if organisations use productivity gains solely to demand more output. 

The consequences extend beyond employee well-being. When work becomes more fragmented and intense, employees may find it harder to think deeply, generate novel ideas and solve complex problems. Burnout also contributes to turnover, often leading organisations to lose experienced staff whose judgment, context and institutional knowledge cannot easily be replaced by AI. 

The risk is that companies produce more output while reducing their capacity for good decision-making and effective execution. Sustained performance rarely improves when employees are chronically exhausted. Over time, burnout erodes engagement, weakens retention and undermines results.

Accepting human limitations 

Burnout is not an inevitable consequence of the AI revolution. But avoiding it requires organisations to acknowledge human limits and create conditions for people to flourish rather than assuming that every productivity gain should be converted into additional work. 

First, leaders need to recognise that AI increases the importance of human recognition rather than diminishing it. Employees need reassurance that judgment, experience, creativity and ethical reasoning remain valuable. 

For example, one executive told me that they deliberately highlighted to their team occasions when an AI tool produced unrealistic recommendations, and when a human colleague improved the outcome. By doing so, they reinforced the value of human judgment rather than allowing the technology to take all the credit. 

Second, leaders should resist the temptation to convert every efficiency gain into higher workloads. In some cases, a portion of the capacity created by AI should be returned to employees rather than immediately redirected into additional output. 

Leaders should remember that long-term performance depends on human sustainability. The organisations that benefit most from AI will not necessarily be those that push people the hardest. They will be those that create the conditions for people to exercise judgment, build relationships and generate new ideas – capabilities that remain difficult to sustain in a state of chronic exhaustion.