Ekaterina Hertog is an associate professor of AI and Society at the Oxford Internet Institute, and leads the research project DomesticAI. As anxieties over artificial intelligence replacing jobs reach a fever pitch, Hertog has been asking how a country like Japan can benefit from automation in the home.

This interview has been edited for length and clarity.

Why were you drawn to studying the impact of AI on domestic work?

We hear from academics, but also from the media, about the future of paid work. “These jobs will disappear.” “Who is particularly at risk?” My research was very much focused on unpaid domestic labor, which is housework and care work. It’s a necessary type of work that has to be done just as much as paid work for societies to function.

How is automation changing domestic work now?

All of us are familiar with the vacuum cleaner. In more recent years, there is a lot of discussion about smart baby rockers. It’s able to adjust rocking a baby, eventually finding out the way that particular baby enjoys being rocked to pacify them. I’ve seen a number of promotions for a so-called laundroid machine developed by Panasonic. It looks like a really big wardrobe with various drawers — you throw in your dirty laundry and eventually get it out clean and ironed.

What kinds of domestic work are most primed for automation?

We asked experts about 17 types of domestic tasks: cooking, dishwashing, cleaning, grocery shopping, and physical child-care like changing nappies, helping with homework. Across all these categories, the lowest prediction in five years’ time said 20% of time could be saved through automation. The highest was specifically grocery shopping. There’s something called ‘predictive shopping’: It sees your shopping patterns, and then sends you things you want or need without you actually having to order them. You can return whatever you don’t need.

Why is there a particular demand for automating domestic work in Japan?

Women in Japan still continue to do almost 90% of all housework and care work within couples. So when they are married, they do not just do the majority, they do almost everything. That has major consequences on their ability to earn.