What if robots did all the chores?

Chores are jobs at home, like sweeping, washing dishes and folding clothes.

What if robots did all of them? What could robots do, and what would be hard?

Every home has chores: vacuuming, washing dishes, doing laundry, tidying up. Machines already help with some of them. Dishwashers and washing machines do a lot of the work, and small robot vacuums roam floors on their own.

So what if robots did all the chores? This is a thought experiment about real technology. We'll look at what household robots can do now, what they still find surprisingly hard, and what more robot help could mean for people.

Household robotics sits at the boundary between what automation does well and what remains difficult. A floor is a predictable, mostly flat surface; a pile of laundry is not.

This thought experiment asks what it would take for robots to do every chore, why some tasks resist automation, and how wider automation affects work. It draws on the International Federation of Robotics, university robotics research and MIT's Task Force on the Work of the Future, and labels fact and idea.

Robot vacuums are little round robots. They roll around and clean the floor by themselves.

Some can mop too. Many can make a map of the rooms!

Robot vacuums are the most common robots in homes. The International Federation of Robotics (IFR) reports that more than 2 million floor-cleaning robots were sold in 2023 alone, about 57 percent of all household service robots.

They've come a long way. Early robot vacuums moved around at random. Today most can wet mop as well as vacuum, and they use cameras and smart navigation to map rooms in 3-D, so they clean more efficiently.

IFR data show floor-cleaning robots account for close to 57 percent of domestic service robots, with more than 2 million units sold in 2023; growth has slowed since a pandemic-era peak, and IFR calls the market fully mature. In 2001 about 5,000 units sold in Sweden; by 2004, IFR estimated 570,000 sold in total.

Early models navigated by random walk. Most current models also wet mop and use systematic navigation and cameras for 3-D mapping.

If robots did every chore, people might have more free time. That sounds fun!

But robots would need to fold clothes, tidy toys and much more. Those jobs are hard for robots.

Imagine coming home to folded laundry, a tidy room and clean dishes, all done by robots. People might have more time for other things. That's an idea, not a promise.

To get there, robots would need to master jobs that are easy for us but hard for them: picking up toys of every shape, folding clothes, putting things where they belong. Researchers are working on these skills, but today's robots are still slow and often struggle with even simple folding.

Full household automation would require reliable manipulation of unpredictable, deformable and varied objects, perception in cluttered spaces, and safe operation around people. Knowable Magazine reports that current folding robots are slow and often perform poorly on even simple tasks.

What universal chore robots would change, in free time, costs or habits, is speculative. These consequences are ideas, not forecasts.

A floor stays flat. But a T-shirt can scrunch up in lots of different ways.

Every scrunch looks new to a robot. That makes folding very tricky!

Factory robots can build cars with amazing precision, but only in carefully controlled settings where everything is in the same place every time. Homes are messier. Berkeley researchers explain that clothes are 'deformable': floppy, so their shape isn't predictable.

Carnegie Mellon robotics researcher David Held says the challenge isn't the fabric itself but how many ways it can crumple. People also have flexible, sensitive hands, while many robot grippers are built for one shape. A 2024 study found a standard robot folding method scored only 0.41 out of 1 for a neat fold; a newer method that adjusts as it folds scored 0.83.

Industrial robots perform with great precision in highly structured settings but are much more limited outside them (Maitin-Shepard and Abbeel, Berkeley). Deformable objects defeat vision methods designed for rigid ones, and every crumpled configuration is novel.

'Pick and place' folding uses predetermined motions that fail when fabric shifts; adaptive approaches re-plan during the fold. In a 2024 study, pick and place scored an intersection-over-union of 0.41 folding a cloth in half, versus 0.83 for AdaFold. Danica Kragic (KTH) notes fabric handling needs both advanced manipulation and high-level reasoning.

When machines take over some jobs, people often find new jobs to do.

People design robots, build them and fix them!

What happens to people's work when machines do more? MIT's Task Force on the Work of the Future found that technology replaces some jobs while creating new ones. About 63 percent of the jobs people did in 2018 didn't even exist in 1940.

The researchers describe a gradual change, not a sudden robot takeover, and say good choices about how technology is designed and used help decide who benefits. So a world with more robot helpers would likely bring new kinds of work too. Exactly what kinds is something nobody can know yet.

The MIT task force's 2020 report concludes that technological change simultaneously replaces existing work and creates new work; about 63 percent of jobs performed in 2018 did not exist in 1940. It describes 'a gradual tech evolution' rather than a robot revolution, while noting that the gains have not been shared evenly.

Its leaders emphasize that outcomes depend on human choices in design and policy. Which household-robot jobs would emerge is speculative; they might include roles in designing, building and repairing robots.