Women Still Do More Housework Than Men. Can AI Help?

09/19/2023
By Ed Brennen

Since ChatGPT launched late last year, countless news stories, podcasts and think pieces have examined the potential impact of artificial intelligence, or AI, on education and the workforce.

One area that has received far less attention, though, is AI’s impact on life at home.

In a newly published research paper for the Brookings Institution, a policy think tank, Manning School of Business Assoc. Profs. of Management Elizabeth J. Altman and Beth Humberd explore how AI and related emerging technologies like autonomous vehicles can help with the day-to-day tasks and responsibilities of running a household – and whether the technology can help alleviate the age-old disparity between men and women when it comes to taking care of this work.

“The better these technologies become, the more they will help people who are managing families, children and households,” Altman says. “The question is: How does that affect the conversation about domestic labor?”

Altman, who was recently named a Nonresident Fellow in the Economic Studies program at Brookings, co-authored a policy report for the institution in April on workforce ecosystems and AI. When she was asked to write a follow-up paper on the promise of emerging technologies to address the gender gap in domestic labor, she brought in Humberd, whose research expertise includes dual-career families and the changing nature of work.

“We've worked alongside each other for a long time but never collaborated on a paper, so it was a nice intersection of both of our areas of expertise,” Humberd says.

Altman and Humberd sat down to discuss the research, which focuses specifically on married or cohabiting heterosexual couples in the United States, with or without children – but could be expanded in the future to include other family environments.

Q: How big is the gender gap in household labor?

Humberd: Fifty years ago, of course, women did more of the housework and men did more of the paid work. As more women entered the paid workforce and the number of dual-career couples increased, men started to pick up more of the slack and take on more household labor and child care. But the gap persists today, even in dual-career couples and, most interestingly, even where the woman is the chief breadwinner. The Bureau of Labor Statistics does American Time-Use surveys that show women are doing about one and a half times as much as men when it comes to household management and caregiving – about 4.5 hours to 2.75 hours per day.

Q: You differentiate between time spent “doing” domestic labor like cooking and cleaning and the mental and emotional energy required for “managing” things like grocery lists and doctor’s appointments. Why is that important?

Altman: Time-use surveys fail to fully capture the invisible components of managing domestic labor – the anticipating, planning, deciding and overseeing of household tasks. When you think about how technology can do something faster and more efficiently, the notion of that efficiency resulting in a reduction in cognitive load hasn’t been discussed in the domestic sphere. It's increasingly being discussed in the workplace domain – you hear metaphors of co-pilots and having a team of assistants around you – but it hasn’t emerged as a big theme with household labor. We think it will.

Humberd: Prior technological advancements in the domestic sphere – we went back and looked at when the washing machine was invented – were often framed as helping because they were time-savers. Yes, AI and autonomous vehicles are likely to provide some time-saving, but we think that the unique promise is in reducing some of the cognitive and mental load. Personally, I’m always multitasking, thinking of the “other thing.” If I have a break in between classes, I’m at work thinking, “Oh, let me schedule that doctor’s appointment for my kids’ annual physicals.” AI might not help with the inequality of hours, but more with that mental load.

Q: You give examples of platforms and technologies that already help reduce this mental load: Amazon alerting you when it’s time to reorder a product, HelloFresh sending you meal kits, health care apps that enable you to manage appointments. How will AI enhance this? 

Humberd: ChatGPT makes AI more accessible. We are just at the start of the regular layperson even touching this technology. Who knows where it's going to be a decade from now? As the technology improves, HelloFresh will learn even more about what your family likes. There is kind of a sweet spot here where something like ChatGPT could perhaps scrub your emails for things like soccer games and dentist appointments and make a calendar for your entire family.

Altman: I'm already linking calendars with family members, but it's still quite manual. Could having my calendar sync in a smart way that catches conflicts be much better in the next 18 months? Yes. A lot of these systems are becoming more user-friendly. The computing power needed to run them is becoming much more accessible, both in our own devices and in the cloud. Plenty of these systems existed and could do this before, but they couldn't do them for every 12-year-old with a smartphone.

Q: The hype and promise of self-driving cars seems to have stalled a bit. How do you see autonomous vehicles contributing to domestic labor?

Altman: We still may be a long way from “The Jetsons,” but the technology is continuing to improve and become more and more useful. Waymo, for instance, recently introduced a driverless taxi service in San Francisco. I don’t think most people are ready to put their child in a self-driving car yet, but maybe eventually they will. But are they OK with an autonomous vehicle pulling into their driveway to bring them dinner? Perhaps. 

Q: When someone uses this technology, you highlight the “second-order effects” it has on others in the household. What are they? And why is that significant?

Altman: People are very focused on what this technology can do and how it can change what the people using it do. How the evolution of this technology affects broader systems of humans, people connected to technology users, is also very interesting. In our particular study we focus on family units. How the technologies affect one or more of the members since they are all connected to each other, is a broader question. It seems obvious after the fact, but to me that was a pretty strong insight.

Humberd: We give an example of a father who used ChatGPT to find things to do with his young son on a rainy day while his wife was at work. They decided to bake cookies but lacked some of the ingredients. Rather than calling mom and asking her what to do, they went back to ChatGPT and figured it out. So the technology wasn’t reducing the woman’s workload directly, but she benefited from the fact that her family was using it at home.

Q: What are the policy implications of your research?

Altman: There are big questions around equity – ensuring more vulnerable populations like the elderly and disabled have access to new technologies. With autonomous vehicles, there is the question of risk and safety. And then there are the privacy questions around data – not only the data that people provide willingly, but also that systems can access through surveillance. All those questions play into policy considerations. We are just at the tip of the iceberg of learning about them, but it’s a great opportunity to combine business and management research with technology research to learn more about policy.

Humberd: There needs to be transparent collaboration between the companies building these technologies and government regulators like the Federal Trade Commission. As consumers, I don't know that we're capable of managing the way these technologies exist in the home. … But I think it’s exciting to see how AI could be used to solve an age-old problem. I don't think there's going to be some AI that shows up next week that ensures there’s never going to be an unequal division of labor again. But maybe it can poke holes in this sticky problem.