A slight twist of the wrist reveals a cursor on the screen. Tapping with the thumb and index finger opens the app. Writing letters in the air with a pencil displays the characters on the screen. A wristband that controls computers, laptops, and smartphones with hand gestures has been released. It can control devices both right next to it and in the room across.
Meta, the U.S. big tech company that operates Facebook, Instagram, and WhatsApp, noted in a study published on the 24th in the international journal "Nature" that it has developed a wristband that reads the electrical signals flowing in finger muscles to anticipate movement intentions and control devices.
Before people act, they first think about how they want to move in their minds. Then, the brain sends motor commands to the muscles in the form of electrical signals. The wristband developed by Meta Reality Labs reads these electrical signals flowing from the arm muscles.
The electrical signals sent by the brain are very powerful and can be read even outside the skin. They move much faster than muscles, particularly. By reading the electrical signals in advance, it would theoretically be possible to enter text faster than a finger.
The research team trained artificial intelligence (AI) to identify the electrical signals that emerge when a person moves their fingers and wrists. The AI learned from data collected from 10,000 individuals who tested the wristband prototype, mastering the different electrical signals corresponding to various hand movements. Later, the AI could detect the motions of individuals wearing the wristband for the first time and control the device in the same manner.
The wristband controls smartphones, laptops, and computers in real-time via Bluetooth. Currently, it can detect gestures at a rate of 0.88 times per second. The speed of inputting text into a smartphone using hand gestures is 20.9 words per minute, slightly slower than the actual finger speed of 36 words.
Meta's research team explained that with practice, users could input text much faster than with fingers. Furthermore, they projected that it could detect desires simply by thinking without actually moving hands. Thomas Reardon, Vice President of Research at Meta, told the New York Times that "with just a bit more practice, there will be no need to move hands at all" and that "just thinking about wanting to move will allow you to control the mouse cursor."
Previously, scientists have developed technologies that allow users to control devices without touching them. A representative example is the Brain-Computer Interface (BCI) chip developed by Neuralink, a startup founded by Elon Musk, and Synchron from Australia.
BCI is a technology that converts brain waves into electrical signals, allowing information exchange with computers. However, implanting small chips under the skull or in neck vessels was hazardous and cumbersome. Therefore, clinical trials were only conducted on patients with limb paralysis. Additionally, because it is based on directly decoding brain activity, there is a limitation that it takes time for users to become accustomed.
Meta's wristband, which detects electrical signals flowing in muscles rather than the brain, has been deemed much simpler and safer than BCI chips. Dario Farina, a professor in the Department of Bioengineering at Imperial College London, evaluated that "the remarkable aspect is that Meta has strengthened this technology by analyzing the big data of prototype users using AI."
Meta announced that it plans to release this wristband product within a few years. Last autumn, Meta demonstrated capturing photos and videos and playing music using the wristband. Additionally, it is currently collaborating with researchers at Carnegie Mellon University in the U.S. to test the wristband on patients with spinal cord injuries.
Douglas Weber, a professor in the Department of Mechanical Engineering and Neuroscience at Carnegie Mellon University, explained that "because it reads the intention to move before the action, patients with amputated fingers, weakened muscle strength, or limb paralysis can also control the device."
References
Nature (2025), DOI: https://doi.org/10.1038/s41586-025-09255-w