AI made the shift impossible to ignore — but platforms, algorithms, feeds, and devices have been doing it for years. The evidence is scattered across fields. The research has rarely been connected.
The Conduction Lab develops measurement frameworks, detection tools, and governance architectures for a phenomenon observed across every domain but studied in none as a unified problem: the systematic shift of human baselines by the systems people interact with.
Our foundational research identifies the mechanism. Large language models don't generate human-like signal — they conduct real human signal through training data, producing systematic source misattribution at the point of reception. This is the Conduction Hypothesis, and it reframes how we study human-AI interaction, AI safety, and the broader question of what happens when systems mediate human experience.
The methodology is native to pharmacovigilance and clinical research — signal detection, causality assessment, and baseline measurement applied to a new exposure class. The work extends beyond AI to every system that shifts a baseline: social media, devices, relational dynamics, and the body itself.
The theoretical foundation. The Conduction Hypothesis, misconduction as a failure mode, corpus dynamics as the study of how humans interact with the accumulated body of what they've received and stored, and the frameworks that make baseline shifts detectable.
ExploreTheory turned into infrastructure. Detection frameworks, authentication protocols, clinical AI governance tools, and product architectures that operationalize the research into systems people and organizations can use.
ExploreAcademic presentations, submitted research, and long-form writing. Peer-reviewed work alongside public writing that develops ideas before formal submission.
ExploreThe Conduction Lab was founded by Kim Hosein, MBA — a clinical project manager in pharmaceutical research whose work in pharmacovigilance and clinical operations provided the native methodology for studying what AI systems actually do to the people who use them.
The lab applies signal detection and causality assessment — methods built for identifying adverse events in black-box systems where mechanism is invisible and attribution is assessed from output — to the question of how systems shift human baselines. The result is a research program that doesn't require access to model internals to produce actionable findings.