Brain–Machine Interfaces (BMIs) are rapidly transitioning from controlled laboratory environments into wearable, semi-autonomous systems supported by consumer devices such as smartphones and laptops. While significant research has focused on signal acquisition, decoding accuracy, and power efficiency, comparatively little attention has been paid to operating-system–level security models tailored specifically for BMIs—systems that directly interface with human cognition and intent. This paper introduces THEMI-OS (Thermal-Energy Minimal Interface Operating System), a minimalist, security-first operating system designed for non-invasive, energy-harvesting BMIs that rely on external host devices for computation and feedback. THEMI-OS adopts a brain-sovereign trust model, emphasizing intent abstraction, asymmetric trust, ephemeral communication, and reflexive failure modes. We present its architecture, intended use cases, benefits, and limitations, and critically examine contexts in which THEMI-OS should not be deployed. The goal is to offer a balanced, credible framework that advances BMI safety without overstating feasibility or universality.
Krishna Mohan Avancha (Fri,) studied this question.