Meta has made a groundbreaking leap in brain-computer interface (BCI) research with a device that can translate your thoughts into text. Using a powerful brain scanner and deep learning algorithms, Meta’s team has demonstrated the ability to reconstruct entire sentences based solely on the neural signals generated while a person types. However, despite its impressive potential, this technology is far from hitting the market due to its massive size, complexity, and cost.
Meta’s innovation involves a magnetoencephalography (MEG) scanner, which detects magnetic signals in the brain without the need for invasive implants. This non-invasive approach has drawn attention as it could allow users to “type” by simply thinking about what they want to say. According to the research, Meta’s deep-learning system, known as Brain2Qwerty, was able to predict keystrokes with up to 80% accuracy — a major step forward in mind-typing technology.
Yet, the device is not without its challenges. The MEG scanner, which weighs about half a ton and costs around $2 million, requires a shielded room to function properly. The user must also remain perfectly still during use, which limits its practicality. This makes it highly unlikely that the technology will become commercially viable anytime soon.
Despite these hurdles, Meta sees valuable insights from this research that could advance AI development. Jean-Rémi King, leader of Meta’s Brain & AI team, believes the findings could inform the future of artificial intelligence, particularly in understanding how language is processed by the human brain. “Language has become a foundation of AI,” King stated, adding that the principles behind the brain’s language formation could shape the next wave of AI technology.
