JudgeAI

JudgeAI explores a single question: how can AI make decisions about rules and actions? Its approach is autonomous normative decision-making, a process in which the system, not a human, arrives at the decision. It begins with an ontology — legally significant states and relationships compressed from the real world into a bounded, computable structure.

From there, every mutually exclusive course of action, or inaction, is projected into a different possible state of reality. JudgeAI compares how consequences propagate between participants under one consistent normative framework, then returns a result. That output can be a draft bill, a draft judicial decision, or an instruction to an AI agent, and each comes with traceable grounds that feed further research and recalculation.

The project positions itself as an ongoing research program rather than a packaged product, with a long-term vision of a shared normative layer that every resource-controlling AI agent acts through. Alongside the research it publishes a Legislative Atlas with RSS and JSON feeds spanning jurisdictions such as the UAE, Ireland, the UK, the Netherlands, the US and China. Access is arranged through a scheduled call rather than a public pricing page.