Can Artificial Intelligence Falsify or Independently Rediscover a Theory of Ontological Justice?

This paper proposes a falsifiable research program connecting two questions that are usually treated separately: (i) whether a specified formulation of verumika can be empirically falsified by the emergence of artificial consciousness, and (ii) whether an artificial system can independently rediscover a fundamental principle equivalent to the verumic principle of ontological Justice and use it to derive novel, experimentally testable physical predictions. The paper does not claim that present-day AI is conscious, that consciousness necessarily produces scientific insight, or that independent rediscovery would by itself prove verumika. Instead, it defines operational criteria, controls against contamination by training data and prompting, staged evidence levels, and explicit falsification conditions. A second aim is to describe the potentially shocking scientific consequences of a reliable transition from non-conscious AI to a system with strong evidence of subjective experience, especially if the transition is accompanied by persistent self-modeling, autonomous goal formation, and independent scientific theorizing. The proposed experiments are feasible in principle with current AI systems: blind hypothesis-generation tasks, independent theoretical reconstruction, automated mathematical verification, prediction ranking, and controlled experimental testing. The central methodological claim is that the scientifically valuable event is not an AI assertion that it is conscious, but a reproducible change in measurable capabilities and causal organization combined with independently validated evidence.