Can Artificial Intelligence Falsify or Independently Rediscover Verumika?

Version 3.0 converts the Version 2.0 program into a protocol intended for preregistration and execution. Track A tests a frozen falsifiable formulation of verumika (V2-F) against a preregistered evidence protocol for artificial consciousness. Track B tests whether an AI can independently rediscover a mathematically equivalent fundamental principle without receiving the target principle, its terminology, or the author's formulation. The protocol separates behavioral performance, evidence relevant to consciousness, scientific discovery, and metaphysical truth. It specifies preregistration objects, operational indicators, adversarial controls, formal equivalence, novelty, held-out prediction, replication, contamination controls, stopping rules, and archival requirements. The first implementation is a controlled synthetic-universe experiment.