What happened in this episode?
At a Summit 2026 session, moderator Rana El Cahibi interviews Matt Angel, CEO of brain-computer interface company Paradromics, and Duke University professor Nita Farahany about the technical trajectory and governance risks of brain implants. Angel presents video from what he describes as Paradromics’ first clinical trial involving a woman with motor neurone disease, saying the implant decodes words from her “inner monologue” into on-screen text, and he explains the device records from a language-related area he calls “55B”.
The conversation pivots from assistive communication to consumer neurotechnology and the question of who controls neural data. Angel argues that trust should be built around business models as well as individuals, suggesting medical-device reimbursement aligns incentives towards patients. Farahany counters that disputes over ownership are already emerging, alleging some patients have been told they do not own their neural recordings and warning that “offers on the table” to buy neural data could drive commodification.
Both discuss the convergence of BCIs with AI models used to interpret neural signals. Angel describes an “engram” small language model that can revise predicted words as more brain data arrives, and he argues higher data-rate interfaces preserve user agency by reducing how much AI must “fill in”. Farahany warns that earlier-stage brain signals could capture bias or private thoughts and that AI-mediated outputs may be experienced as self-generated, motivating her push for “cognitive liberty” and design standards centred on user self-determination.