The commentary reviews recent milestones, including company-reported Neuralink implants in patients with severe paralysis (12 as of September 2025; a stated target of 20,000 implants annually by 2031), early human testing of China’s semi-invasive “BeiNao-1” epidural system (five patients by early 2026, as cited), and Stanford’s inner-speech decoding study published in Cell in August 2025 (four participants; reported >98% accuracy in a passcode-protection/authentication task). These figures are reported by the groups or in the cited literature and should not be read as independently validated clinical outcomes. Technically, invasive BCIs face glial scarring and signal decay; decoding remains limited by “black box” models and poor generalization; and a “bandwidth chasm”—a mismatch between available neural signal and the speed and richness needed for natural control—may limit real-time control and user agency. Ethically, the authors highlight neural privacy, citing a 2023 Chilean Supreme Court ruling concerning Emotiv and evidence that facial-recognition algorithms identified 83% of participants from de-identified MRI scans (as cited in the commentary). They warn of a “neuro-social divide,” regulatory gray zones for consumer neurotechnology, and challenges of informed consent in cognitively impaired patients. The authors call for precautionary, stewardship-oriented governance prioritizing human dignity, safety, and equitable access alongside innovation.

This work, entitled “Brain-Computer Interfaces at a Crossroads: A Critical Analysis of Technical, Ethical, and Societal Hurdles on the Path to a Connected Future,” was published on Human Brain(published on Sep. 28, 2026).
DOI:10.15302/HB.2026.0001