However, there is a new frontier to express hope. Tech giants have continuously collected and started using data on human movement. They are at the forefront of the race to crack the code and accelerate the development of humanoid robots that interact flawlessly with human life. It focuses on one of the main aspects: teaching AI not only to think and respond appropriately in the virtual world but also to master the art of physical movement in everyday settings. Therefore, companies in the US, China, and other regions are redoubling their efforts by developing AI using videos and sensor data that provide high-precision insights into how humans perform.
For example, a startup in India hires people to perform monotonous, well-supervised jobs, like folding towels, while wearing cameras that record precise movements. These films and data then serve as a basis for teaching robots to do the same. Moreover, other businesses use teleoperation, in which humans control enslaved people, which helps AI improve performance and problem-solving understanding by observing successful and failed actions. This occurs because a single dispatch controls teams across the ocean, people drive the robot with joysticks, and coordinated robots learn to act quickly in different environments.
A few of the leading companies, notably Meta-backed Scale AI and San José-based Figure AI, have raked in up to $1 billion to harvest first-person human movement data in our homes and workplaces. In these companies, people’s behavior is being scrutinized —from how we walk to how we reach for objects. The ambition is for robots to move across and interact with such human spaces seamlessly and safely. This category of data collection helps robots improve motion control, perform demanding activities such as cooking, cleaning, and tool manipulation, and demonstrate dexterity and adaptability.
Still, the complete autonomy of humanoid robots cannot be achieved: the majority of the equipment available today relies on human commands for movement guidelines and can be used only in a controlled environment, primarily in factories and warehouses. Although the development of artificial general intelligence and sensory innovations enhances the perceptual and cognitive aspects of robots, while also improving movements with generative and responsive forced perfection systems, the robot units are still powered by human-made motor systems; therefore, the duration of operation is relatively brief, and intricate labor cannot be done.
With the help of new technologies, robots can learn to optimize their movements. For example, in reinforcement learning, robots will be rewarded when they successfully copy a human movement. Moreover, new motion-capture systems help robots more naturally copy human movements, so that a future humanoid can perform tasks without the need for reprogramming. In other words, future humanoids are likely to adjust to new situations in real life by abandoning their rigid, preset behaviour.
The global effort to create humanoid robots that move and work exactly like humans has attracted significant interest and investment, with several major US and Chinese tech giants involved. Analysts estimate the market for humanoid robots will grow to tens of billions of dollars over the next 10 years, driven by software for manufacturing devices and factory procedures, as well as personal butlers, healthcare, customer service, and smartphones.
However, they also acknowledge that companies and governments face various complex ethical, social, and economic issues as more humanoids are deployed. As companies roll out these technologies, they must develop new policies regarding their impact on the workforce, safety, and the proper interaction between humans and robots. Nevertheless, the hope that humanoids will complement human work and enhance safety in perilous situations is driving further research and testing worldwide. To sum up, the human race’s quest to obtain human movement data for training humanoid robots is a critical new frontier in AI and robotics. By tapping into vast datasets on human behavior, the tech giants are expanding the scope of robots’ abilities, endeavoring to produce robots that not only reason, learn, and act like humans but also augment them, enabling them to travel and work effortlessly in the physical world.