Your brain activity can now turn into text without typing:Meta unveils ‘Brain2Qwerty’ AI system that will help patients with neurological disorders

Imagine writing with just your thoughts or what if you could type a message without touching a keyboard or even speaking? That idea may sound like science fiction, but Meta has taken a major step towards making it possible. The company has introduced ‘Brain2Qwerty v2,’ a new artificial intelligence (AI) system that can convert brain activity into text without requiring any surgical implant. Meta says it is the most accurate non-invasive brain-to-text system developed so far and could one day help millions of people who are unable to communicate due to neurological conditions. What is Brain2Qwerty v2? Brain2Qwerty v2 is an AI-powered brain-computer interface (BCI) that translates brain activity into written text. Unlike many advanced brain-computer systems that require electrodes to be surgically implanted in the brain, this technology works without any surgery. Instead, it uses magnetoencephalography (MEG), a machine that records the brain’s magnetic signals from outside the head. AI then analyses these signals and converts them into words and sentences.

Meta described the breakthrough by saying: Brain2Qwerty v2, the highest-performing end-to-end pipeline capable of real-time sentence decoding from non-invasive brain recordings, approaching levels of accuracy previously exclusive to techniques that require brain surgery. How does it work? To build the system, Meta trained the AI using around 22,000 sentences collected from nine volunteers. Each participant spent nearly 10 hours wearing an MEG scanner while actively typing. Instead of relying on manually designed methods to interpret brain signals, Brain2Qwerty v2 uses end-to-end deep learning, allowing the AI to learn directly from raw brain activity. Meta also fine-tuned large language models (LLMs) on neural data so the system could use language context to better predict sentences, even when brain signals were noisy. Also read: Planning to buy iPhone 18 Pro in September?: Company faces risk of pirated versions hitting market after hackers steal critical data from Tata Electronics

How accurate is it? According to Meta, Brain2Qwerty v2 achieved a 61% word accuracy rate, a significant jump from the roughly 8% accuracy achieved by previous non-invasive brain decoding methods. For the best-performing participant, the AI reached 78% word accuracy, with more than half of the decoded sentences containing one word error or less. The company also found that accuracy improved as more training data became available, suggesting future versions could become even more reliable.
Why is this important? Today’s most accurate brain-computer interfaces usually require surgery to place electrodes inside the brain. Although effective, these procedures are expensive, invasive and not suitable for everyone. Brain2Qwerty v2 shows that AI can achieve much better results without surgery, making brain-computer interfaces safer and potentially more accessible in the future. If further developed, the technology could help people affected by conditions such as strokes, brain injuries, paralysis and neurological disorders that make speaking or typing difficult. Beyond communication, it could also support neuroscience research by helping scientists better understand how the brain processes language and develop improved treatments for neurological diseases. What does it mean for India?
The technology could be especially valuable for countries like India, where stroke and neurological disorders affect millions of people every year. If Brain2Qwerty or similar systems become clinically approved in the future, they could help patients who have lost the ability to communicate due to brain injuries or diseases. India’s growing AI and healthcare research ecosystem, including institutions like the Indian Institute of Science, Indian Institutes of Technology and All India Institute of Medical Sciences, could also benefit from Meta’s open-source research to advance brain-computer interface technology.
Also read: Using Netflix through someone else’s subscription?: New login rule now requires a unique email ID for every profile; here what you should do

Still a long way to go
Despite the breakthrough, Brain2Qwerty v2 is not ready for everyday use. The system currently depends on large and expensive MEG scanners and has only been tested on a small number of participants. Even so, the research represents a major step forward in brain-computer interfaces. If accuracy continues to improve, AI-powered systems like Brain2Qwerty could eventually provide a safer alternative to surgical implants and give a voice to people who cannot communicate through speech or movement.