Summer holidays are approaching, and with them comes our craving for the freedom to finally do what we actually want to do.

Perhaps that is also why, when people around the world are asked what excites them most about AI’s potential, the answer most frequently given is remarkably simple: it could save us time.

The promise feels familiar.

Throughout history, nearly every major technological innovation arrived with the expectation that it would reduce effort and create more leisure. The washing machine, the vacuum cleaner and the dishwasher all carried the promise of freeing us from labour.

Yet historically, something more paradoxical often happened instead: rather than doing less, we started doing more as our expectations changed as well with each big technological innovation.

As sociologist Ruth Schwartz Cowan famously argued in More Work for Mother (1983), household technologies did not reduce the total amount of domestic labour. Standards simply increased. Clothes were washed more frequently. Children changed outfits daily. Men stopped wearing detachable collars and cuffs, meaning entire shirts now needed laundering.

In a recent essay for The New York Times, economist Carl Benedikt Frey, professor at the University of Oxford and author of How Progress Ends: Technology, Innovation and the Fate of Nations, adds another perspective to this paradox. AI may not simply take over work and free up our time. Instead, it increasingly enables us to do the work ourselves.

Let’s take the washing machine as an example once more.

Before households adopted washing machines en masse, laundry was often outsourced to professional laundresses. Washing clothes was incredibly labour-intensive. As Frey writes, “Water had to be hauled, fuel chopped, linens boiled in lye, garments scrubbed by hand against washboards, wrung out, dried, starched and ironed with heavy flat irons heated on a stove.” Doing the laundry for a household could easily consume an entire week.

The washing machine undeniably made the work easier. But rather than eliminating the labour altogether, it shifted the labour into the household itself. Families stopped paying others to do the laundry and began doing it themselves.

The laundress lost her job.
The housewife gained an extra chore.

This pattern has repeated itself ever since.

Self-checkout turns shoppers into cashiers. Online booking platforms replaced the travel agent. Online brokerages transformed us into our own bankers. And now AI is accelerating this self-service society by extending its logic into fields that once required trained professionals: healthcare, law, administration, finance and even emotional support.

As of January this year, more than 40 million people worldwide were reportedly using ChatGPT daily for health-related questions, from checking symptoms to decoding medical bills and fighting insurance claims.

This shift towards the self-service society initially feels empowering. We gain access, speed and independence. We save money. We no longer need to wait for the expert.

But something may also quietly disappear in the process.
Because self-service does not automatically reproduce professional judgment.

The accountant points out the deduction you did not know existed. The doctor notices the symptom you overlooked. The strategist recognizes the cultural shift you did not yet realize was happening.

Frey describes this as the AI trade-off: “greater access, but thinner expertise.” The tool answers the question you ask. The expert understands which questions matter in the first place.

At the same time, another effect emerges. Humans tend to overlook the value of what they give up when the cost comes in time rather than money, a phenomenon economists call “opportunity cost neglect.”

As Frey puts it:

“We notice the accountant’s fee we didn’t pay. We rarely notice the evening we spent doing her job.”

If we all become our own travel agent, researcher, financial advisor, therapist, or socio-cultural trend consultant, assisted by AI at every step, what kinds of professional knowledge slowly disappear? And if technologies designed to save and optimize time continuously end up filling our lives with new forms of unpaid labour, are we actually moving closer to spare time at all?