{"version":1,"type":"rich","provider_name":"Libsyn","provider_url":"https:\/\/www.libsyn.com","height":90,"width":600,"title":"Saturnin Pugnet | Brain-Computer Interfaces in the Age of AI","description":"Saturnin Pugnet is a founding member of World (formerly Worldcoin) and author of Phase Lock, a manifesto exploring the convergence of brain-computer interfaces (BCIs) and artificial intelligence. In this episode, we discuss his argument that advances in AI could dramatically accelerate neurotechnology, particularly through large-scale non-invasive brain recordings, improved decoding algorithms, and growing demand for neural data. We also explore why Sat believes neural data could become valuable for training and aligning future AI systems, whether non-invasive BCIs could eventually scale beyond traditional clinical applications, and what this might mean for human agency and cognitive privacy. Along the way, we discuss the limitations of implantable devices, the economics of neural data, the risks of increasingly readable brain activity, and whether neurotechnology could help humans maintain influence over increasingly capable artificial intelligence. Top five takeaways 1. AI could shift the balance between neural signal quality and data quantity. Traditionally, brain-computer interface research has relied on relatively small datasets, often using invasive recordings to obtain high-quality neural signals. Sat argues that advances in AI could make it possible to extract useful information from much larger, noisier datasets collected using non-invasive technologies. Whether data volume can compensate for the physical limitations of these recordings remains an important open question. 2. Neural data could become a valuable resource for artificial intelligence. Most AI models learn from human-generated artifacts such as text, images, speech, and video rather than directly measured brain activity. Sat suggests that neural recordings could provide additional information about cognition, preferences, and human behavior, potentially becoming valuable training data for future AI systems. This remains a forward-looking hypothesis, particularly regarding its commercial value and contribution to AI alignment. 3. Non-invasive BCIs may have a faster path to widespread adoption. Implantable neural interfaces face significant barriers, including surgery, clinical infrastructure, regulatory requirements, and cost. Non-invasive systems may be easier to distribute and scale, especially for consumer applications that do not make medical claims. The trade-off is that non-invasive technologies generally provide less direct access to neural activity, making their useful decoding capabilities an important technical challenge. 4. Reading brain activity creates new questions about cognitive privacy. As neurotechnology improves, neural data could reveal increasingly sensitive information about attention, intentions, or mental states. Sat raises concerns about surveillance, behavioral manipulation, and AI-generated content that adapts to a user's neural responses. These possibilities create a need for privacy protections, informed consent, and ethical standards as brain interfaces become more capable. 5. Neurotechnology could become part of the AI alignment discussion. Sat proposes that information about human cognition and values obtained through neural interfaces could help steer increasingly capable AI systems toward human interests. The central challenge is whether brain recordings can meaningfully reveal values and preferences, and whether those signals would actually improve AI behavior. The idea remains speculative but introduces an unconventional direction for future neurotechnology research. Timestamps 0:30 Introducing Saturnin Pugnet and Phase Lock 2:30 What is the central argument of Phase Lock? 6:30 Why are AI and neurotechnology converging now? 8:30 AI alignment, safety risks, and the future of intelligence 10:30 How much could AI change the world over the next decade? 14:00 What should neural engineers focus on in the AI era? 15:30 From small invasive datasets to large-scale wearable recordings 18:00 Why neural data could become valuable to AI companies 21:30 Could neural datasets attract hundred-million-dollar investments? 24:00 Regulatory and hardware barriers to scaling BCI 27:00 Human agency, cognitive privacy, and manipulation risks 30:30 Should neurotechnology development be slowed or regulated? 31:00 Where would Sat invest a billion dollars in neurotechnology? 33:30 What does Phase Lock mean?     ","author_name":"Neural Implant podcast - the people behind Brain-Machine Interface revolutions","author_url":"http:\/\/neuralimplantpodcast.com","html":"