Meta’s FAIR research team has made significant progress in developing a non-invasive brain-to-text AI, reducing the average word error rate to 39%. This technology has the potential to revolutionize communication for people with brain injuries or paralysis. The team’s Brain2Qwerty v2 model uses magnetoencephalography to record brain activity and reconstruct sentences with impressive accuracy. With this advancement, the possibility of real-time use is becoming more feasible. Why it matters: this breakthrough could pave the way for more accessible and safer communication solutions, bypassing the need for risky surgical implants and transforming the lives of those who rely on them.