Signal Implementing Computational Law in Wolfram Language for the Governance of Artificial Intelligence

Summary

The paper presents an implementation of computational law for AI governance using the Wolfram Language. The author frames the central question as how to govern AI systems whose internal reasoning cannot be fully inspected. The paper argues that governance does not require understanding a system's internal reasoning, but instead requires stating what the system is obliged, permitted, and forbidden to do, and then checking whether it complies. Following this approach, the paper implements Reified Input Output Logic, the formal system underlying the DAPRECO knowledge base. The implementation includes its core input and output axioms and obligation structures. These are built to run within the Wolfram Language environment.