The TL:DR: Artificial intelligence and large language models, like GPT, are revolutionizing the workforce, impacting a significant portion of US workers’ tasks. While automation may reduce workload, it can lead to unintended consequences such as increased boredom, which has negative implications for health and wellbeing. Effort has intrinsic value, and eliminating it might result in a loss of positive experiences. As automation increases, it’s essential to consider the impact on people’s wellbeing and performance, and be aware of the potential downsides of a less effortful, more efficient life.
Reduced workloads sound attractive
In recent years, artificial intelligence has advanced at an unprecedented rate, with large language models (LLMs) such as Generative Pre-trained Transformers (GPTs) playing a significant role. As highlighted in a recent paper paper, LLMs are poised to reshape the workforce, with roughly 80% of US workers having at least 10% of their tasks impacted. When combined with specialised software and tooling, LLMs have the potential to accelerate 47-56% of all US worker tasks without compromising quality. Earlier research suggests that machines could handle 30% of most job-related tasks, and the use of ChatGPT is associated with a 39% increase in productivity for knowledge work knowledge work . AI is revolutionising the way we work. But what does this mean for our wellbeing and performance?
For the overworked employee, a reduced workload sounds attractive. A robotised future in which humans are on permanent vacation might be idyllic. However, while the human drive to minimise effort and optimise efficiency is understandable and helpful, it may also be associated with unintended consequences.
It’s increasingly clear that our future will be both automated and augmented. Some roles will be entirely redundant. Many will be replaced in part. Complex tasks, once considered too difficult to automate, are being performed by machines. Automated systems generate complicated medical diagnoses and treatment plans; algorithms create detailed, responsive exercise programmes; and artificially intelligent therapists offer patients low-cost programmes to overcome social anxiety.
According to some studies, up to 20% of a senior executive’s role could be automated. AI is likely to be assistive and enhancing, facilitating deeper insights, better decision-making and multiplied output. Work may become less stressful as machines help us to manage information flows more effectively and release us to focus on creativity, collaboration and complex problem-solving. These qualities will be essential in the coming Fourth Industrial Revolution.
Much low-income, manual work will still require human workers. While humanoid robots are becoming increasingly sophisticated, it will take time to robotise these roles entirely. For example, automated vehicles will deliver goods to local hubs. But it will be years until an army of cheap robots is smart enough to navigate the ‘final mile’ through unpredictable entrances, up stairways and into small, rusty letterboxes. However, whether we’re in highly paid knowledge work or more manual roles, there is a risk that more of our work will become supervisory, interspersed with brief periods of activity. Paradoxically, while these shifts may make work less effortful, they might not necessarily make them any easier.
Is reducing effort always a good thing?
Prevailing cognitive psychology, neuroscience, and economic models suggest that mental or physical effort is costly. Given a choice, we prefer to avoid it. In this light, assistive technology which reduces effort may be welcomed. It may make us less stressed and less tired and offer us more free time. A utopian angle might herald a future of automated abundance and mass leisure.
When we consider related historical transformations, automation rarely seems to displace human activity entirely. But it always changes the nature of human work. These changes are often unintended and unanticipated. For example, emerging research draws attention to the links between effort and motivation, cognitive control, value-based decision-making and health conditions. In short, we’re waking up to the intrinsic value of effort.
The effort paradox
Effort can be defined as the subjective intensification of activity – mental or physical – in the service of meeting our goals. As we focus on ways to reduce human effort, we may be overlooking its benefits. Outcomes can be more rewarding if we apply more, rather than less, effort to achieve them. The ‘IKEA effect’ suggests that we will be prepared to pay more for objects we have effortfully built relative to identical objects that someone else built for us.
Effort can also be valuable and rewarding in its own right. Many individuals enjoy cognitive effort for its own sake. ‘Need for cognition’ is a measurable trait associated with an individual placing a high value on mental effort and seeking it out. Recent research sheds light on this phenomenon, helping us to understand why effort can offer intrinsic value. The ‘effort paradox’ explores how the same outcomes can be more rewarding if we apply more, rather than less, effort. It explains how we may select options because they require effort, such as racing a triathlon or climbing a mountain.
We should be cautious about losing the effort habit
As we learn to exert ourselves, we seem able to make more habitual applications of effort over time. Effort is critical in human performance; students show better learning outcomes when their work is effortful. Effort is associated with improved wellbeing, demonstrating positive associations with enhanced goal-directed behaviour: we get better at doing what we aim to do rather than be side-tracked by distraction or temptation.
As we automate more human tasks, we should consider the value of what we are eliminating. What happens if we miss out on positive experiences associated with effort? Will we lose the ‘effort’ habit in the process, with deleterious effects further down the line?
Boredom is halfway between misery and sleepiness
Automation may make boredom – inside and outside the workplace – an increasingly significant issue. This is not a recent observation. At the dawn of the industrial revolution, Nietzsche warned of ‘machine culture’, causing boredom for workers. There are many definitions of boredom, but recent descriptions characterise it as a subjective state of low arousal and dissatisfaction, likely caused by a lack of interest coupled with an inadequately stimulating environment. A study from 1980 put it more succinctly, placing boredom halfway between misery and sleepiness.
While automation may decrease workload and effort for employees, it has been implicated as a source of increasing boredom in some jobs. The boredom risks are often highlighted in safety-critical environments, in which automated systems have increased tedium. In 2009, two pilots were reportedly distracted by their laptops and consequently overflew their destination airport by 90 minutes. However, boredom is already pervasive in more benign office work environments and has been recognised as an essential area for further study.
Boredom may be more fatiguing than effort
In 2017, researchers set out to use electroencephalography (EEG) to monitor the effect of effort and boredom on subjects’ brains. They formulated several hypotheses, predicting that boredom would have a similar impact on the brain to effort. As anticipated, participants exerted greater cognitive effort in an effort condition and felt more bored in the boredom condition. However, while participants in the bored condition initially reported fatigue levels similar to the effort condition, they reported more fatigue as time passed. Boring tasks can be experienced as effortful, and the findings of this study suggest that remaining bored may be more fatiguing than continuously exerting cognitive effort.
Individual susceptibility to boredom varies. Some people can report extreme boredom and others satisfactory interest, even if the environment is identical. Also, workers appear to adopt various techniques to avoid or reduce boredom. More research is needed to determine the most effective means, but approaches to alleviate boredom could be grouped into three categories:
Stimulate: Introduce a secondary, more stimulating task. A machine carrying out 30% of your work may create more time for more creative, non-routine work.
Rest: Schedule tasks so workers get enough breaks to recover from boredom.
Reconsider: Just because something can be automated, should we automate it?
Reconsidering automation
Increasing automation is the end goal for many designers. Many worthy and logical reasons exist for this, including enhanced safety, improved accuracy and decreased costs. However, we should consider the impact on people’s wellbeing and performance. Not only because automation will likely displace human jobs but because the downstream effects of increasing boredom could be severe.
Boredom has been implicated in significant health problems:
– Premature death due to cardiovascular disease.
– Increasing risk of anxiety and depression.
– A reason for recreational drug use in some populations
Recent studies have also highlighted the potential for automation to reduce rather than increase system performance in certain conditions.
Exponential change
The following three trends will likely become more dominant:
Automation: What can be automated probably will be. What cannot will become increasingly valuable.
Augmentation: Those in high-income, cognitively demanding roles will be differentiated by their capacity to work with automated systems that assist and augment their complex work. However, workers in low to middle-income jobs may find that their roles are less effortful but more tedious.
Agility: We are likely to live and work for longer than ever while the world changes around us. We must care for our bodies and minds in a proactive, intelligent way to adapt, continue to learn new skills, and maintain our wellbeing and performance for as long as possible.
While previous transformations have impacted technology, the economy and society linearly, the digital technologies we have created are accelerating development exponentially. Spindles – straight wooden spikes used for spinning fibres such as wool and cotton – took more than a century to spread beyond Europe. It took just seven years for the Internet to cover the planet.
Every technological revolution has brought challenges and opportunities, requiring humans to make difficult decisions and complex judgements. Are feasibility, efficiency, effort and cost savings the most important metrics when considering automation? A less effortful, more efficient life may not be better, particularly if you are bored.