Research

Angus McLeod·2026

An Artificial General Intelligence for Law (AGIL) is a jurisprudentially-grounded neuro-symbolic artificial intelligence (AI) system that produces auditable and accurate legal reasoning. An AGIL has a core of symbolic logic which reasons over formal representations of norms and facts. Neural networks, particularly Large Language Models (LLMs), add scale and flexibility and in return the symbolic core scopes and verifies their inputs and outputs. This symbiotic balance of symbolic accuracy and neural breadth allows for auditable non-monotonic legal reasoning at scale. An AGIL is positivist insofar as what counts as law depends on social facts as normalised across jurisdictions by a weakly universal system design. It is argued that an AGIL of this nature will help to address the

Angus McLeod·2026

Few would describe the experience of being a lawyer in the modern economy as a virtuous pursuit. Since at least the 1970s, the legal profession in liberal societies has been dominated by professional structures, rules, business models and institutions designed to deal with modernity. Those structures, rules, models and institutions have allowed legal professional ethics to extend beyond individual practitioners to firms, outsourced work and now even artificial intelligence (AI) systems. They have also preserved the language of duty and virtue while simultaneously weakening the practices and institutions that cultivate substantive virtue. At one end, modern legal ethics increases moral hazard by primarily relying on appropriate restrictions and punishments to prevent undesirable behaviour,