As humanoid robots are increasingly leaving the laboratory and entering the real world, a completely new category of jobs is emerging – one in which people directly train artificial intelligence through their daily activities. According to a report by CNN, companies developing humanoid robots are increasingly hiring and asking people to record their own household chores, from cleaning to folding laundry, so that data can become the basis for learning robots.
It is a radically different approach to training AI systems. Instead of classic datasets or simulations, real human movement in everyday situations is now used. The equipment required is minimal – often just a camera, smartphone and head mount – but the value of that footage to the industry is enormous.
This work model answers a key question in the development of humanoid robots: how to teach them to function in a world designed for humans. Unlike industrial robots, which work in controlled environments, humanoid robots must understand complex and unpredictable tasks, such as folding clothes or using household appliances. This is exactly where the human example becomes irreplaceable.
The reason for this approach lies in the huge need for data. In order for robots to become “general purpose”, i.e. able to work in shops, offices and homes, it is necessary to collect and process huge amounts of real interactions with the environment.
CNNThis also raises the question of “who” participates in this process. Unlike traditional technology jobs, advanced technical skills are not required here. Practically anyone can become part of the robot training process, which brings this trend closer to the concept of a “data lab”, that is, work based on data generation.
This practice fits into the global race to develop humanoid robotics. Companies like Figure AI, Neura Robotics or 1X Technologies are already developing robots capable of performing household and industrial tasks, but they all face the same problem – the lack of quality data from the real world.
At the same time, the industry is trying to reduce its dependence on humans. Some approaches already use simulations and “world model” systems to help robots learn from their own experiences, but the human factor is still crucial in the early stages of development.
It is important to emphasize that this trend comes at a time when humanoid robots are moving from the experimental phase to commercial application. According to industry analysis, they are already being tested in sectors such as logistics, manufacturing and healthcare, where they can take on jobs that people don’t want or can’t do.
However, it also opens up wider social issues. If people today are training robots that could replace a part of the workforce tomorrow, the question arises of the long-term sustainability of this model. On the one hand, a new type of business is emerging. On the other hand, this same process accelerates automation.
It is precisely in this contradiction that the essence of the current phase of the development of artificial intelligence and robotics lies. People are not only users of technology, but also its key “training set”.
The phenomenon of recording everyday activities for the purposes of training humanoid robots shows how much the boundaries between work and life are being erased in the era of artificial intelligence. What used to be routine is now becoming a valuable resource – information that can help robots one day become our collaborators, but also potential competition on the labor market, reports CNN.