With another round of domestic humanoid robot demos landing in our news feeds, I thought it was worth revisiting a prediction I shared with some friends in November 2024.

My prediction was that domestic humanoid robots will increase the amount of human drudgery in the world, rather than reducing it.

This is less about offering a scenario for dystopian fiction (but hey, knock yourself out) & more about looking at how technology is promised vs how it’s delivered.

The Promise

The promise is that middle to upper-class households will have affordable humanoid robots to do their housework in the very near future. The current pitch is that by combining AI & modern robotics, we’ll have autonomous labour-saving devices that will give people more free time to do things they value & enjoy.

Phase 1: Tele-operated Demonstrations

As demonstrated by companies like Tesla & 1X we have the technology to build humanoid robots with an impressive amount of dexterity. However, so far the demonstrations of these robots have only involved a small amount of AI autonomous movement. The detailed work (serving drinks, doing housework etc.) has been carried out via teleoperation - where a human controls the robot.

Phase 2: Going to Market with Half the Technology

The humanoid robots exist, but the technology for autonomous movement is a long way off. Consider that Tesla has been promising self-driving cars since 2013 & has only just rolled out Robotaxis in limited test areas, with supervising safety drivers in 2025.

So it makes sense to launch the robots, but have them tele-operated. Teleoperation technology already exists & having teleoperated robots in the market will help companies collect real-world data to train their AI. As with most technology, this phase would generally involve early adopters who are happy to shell out thousands of dollars on cool-looking technology with the promise of future functionality.

Phase 3: Growing Awareness & Interest

At this point, we should ask ourselves if there is a market for humanoid domestic robots - beyond well-off early adopters who’ll buy anything shiny.

2.2 million domestic workers are employed in the USA. While some employers enjoy the relationships they have with their domestic help, other households would just like the housework to happen, without encountering the people doing the work.

Having a robot that can do most (if not all) of what domestic help can do, but is available 24/7 without having to manage employment, will appeal to many people.

The first person in a social group who has the novelty of a (tele-operated) robot serving drinks to guests will drive interest amongst their peers.

Phase 4: Tele-operation Develops Faster Than AI Automation

While teleoperation isn’t simple, we're further along & can develop it quicker than AI automation. By its nature our human world is complicated & chaotic. The level of adaptability & dexterity needed to do something as simple as organise a kitchen drawer is beyond current AI systems.

Teleoperation precision can be improved via improved controls, sensors, inputs & network connectivity.

But beyond that, human systems can be put in place for teleoperation, to make sure operators are always available & meeting customer expectations. It doesn’t take much for teleoperation to move from a specialised skill set to something that happens in call-centre style rooms staffed by low-paid workers.

Robots can be limited via software & hardware - for example, not being able to go upstairs 10pm-7am. The human operators can be kept in check through supervision & monitoring of the systems they use to control the robots.

Domestic tele-operated robots always turn up for work, never get sick, take up minimal space when inactive & operators can be switched seemlessly without the customer being aware.

Phase 5: Increased Drudgery vs Freedom From Chores

Rooms full of low-paid workers operating theoretically autonomous technology isn’t new. Amazon’s Just Walk Out technology was supposed to be powered by AI, but relied heavily on manual monitoring by around 1,000 people in India.

Network lag is the only thing limiting a domestic robot in California from being operated by anyone in the world under whatever labour conditions are allowed. From a business perspective, is it more profitable to pour resources into developing AI domestic autonomy, or just outsource the control of the robots to somewhere out of sight & out of mind? Even when AI domestic autonomy is eventually achieved, the business model of tele-operation will be entrenched & probably still outperform AI.

Summary

This is a prediction, not fact (yet). But it’s a prediction worth talking about as it utilises existing technology & business practices to meet a not insignificant market demand.

It’s important that we as society hold companies accountable for what technology they promise, versus how it’s delivered & at what human cost.