Researchers developed a brain-to-voice neuroprosthesis capable of decoding neural activity in real time to synthesize emotive speech in a man with ALS, via 256 microelectrodes implanted in the ventral precentral gyrus. This represents a breakthrough from previous BCIs limited to delayed or text-based communication, enabling continuous speech with prosody, intonation, and real-time modulation. The device’s AI-driven decoding overcame training limitations despite severe dysarthria, offering promising avenues for restoring naturalistic communication to people with paralysis.