On average, 39% of the time currently spent on unpaid domestic work could be automated within the next decade, suggest AI experts from the UK and Japan. The findings are published in the journal PLOS One by a team led by Ekaterina Hertog at the University of Oxford and colleagues in Japan.
A 2013 working paper predicted that automation could threaten almost half of all US employment. This study spoke to growing concerns about “technological unemployment” that arose after the financial crisis, and it was widely discussed in media and policy circles. Following this, other studies were conducted that either criticised or replicated the original study in different countries. These studies also looked into the other effects of automation. The resulting discussions on the “future of work” have captured public attention, informed policy decisions, and directed investments.
According to previous studies, people in the UK aged 15 to 64 spend about 43% of all their work and study time on unpaid domestic work (housework like cooking and cleaning, as well as child or elder care, that could theoretically be delegated to a paid worker or replaced by market goods).
In the UK, working-age men spend around half as much time as working-age women do on such work, and in Japan, the same figure is just 18%. But few studies to date have examined automation in relation to unpaid domestic work, or how predictions about automation differ depending on the AI experts consulted. The authors of the present study asked 29 male and female AI experts from the UK and 36 experts from Japan to estimate how automatable 17 housework and care work tasks might be over the next decade.
The experts predicted that on average 39% of the time that people currently spend on any given domestic work task could be automated within the next ten years. Their estimates varied significantly between tasks, with the most automatable task predicted to be grocery shopping (59%). The least automatable task was physical childcare (21%). UK-based experts believed automation might replace more domestic labour (42%) than Japanese experts (36%). The authors suggest this may be because, in the UK, technology is associated more with labour replacement compared to Japan.
UK male experts tended to be more optimistic about domestic automation compared to UK female experts, which falls in line with previous studies showing that men tend to be more optimistic about technology than women in general. However, this trend was reversed for Japanese experts, with female experts being slightly more optimistic; the authors consider if the Japanese gender disparity in household tasks plays a role in these results.
Though the study’s diverse sample is not statistically representative of the field and is too small to generalize the findings to all AI experts, the authors note that examining experts’ backgrounds may contextualize their forecasting predictions. They also emphasize how these predictions don’t just anticipate the future of work, but also shape it, such that bringing greater cultural and gender diversity to future research is important.
The authors added: “Our study with technology experts in the UK and Japan find that in 10 years’ time, domestic automation could reduce the amount of time spent on current housework and care work tasks by 39%.”
