SYNAPTICON is an artistic-research prototype at the brain–AI model interface, combining non-invasive brain–computer interfaces, foundation models, human–computer interaction, and live multimodal performance. Conceived as a “Brain Waves-to-Natural Language-to-Aesthetics” system, it translates EEG-derived neural activity into language outputs, which then operate as creative indices for immersive audiovisual scenes, storytelling, and altered perceptual experience. Building on this case study, we propose creative alignment as a framework for analyzing expressive neuro-AI systems across signal, semantic, narrative, and social layers. In this view, uncertainty is not merely a technical limitation, but a constitutive condition for co-agency, distributed authorship, and public interpretability. The project contributes a research-through-practice account of how brain–AI interfaces can become responsible artistic apparatuses for novel human expression, while foregrounding ethical questions around cognitive liberty, mental privacy, and the cultural legibility of model-mediated neural data.