Artificial intelligence may be the missing piece that allows legal reasoning to be expressed and calculated with mathematical precision, according to insights shared by Norm AI from its Central Park AI Forum 2026. The question at the heart of the discussion was whether AI can bridge the long-standing divide between natural language and computation.
The session brought together Dr. Stephen Wolfram, CEO of Wolfram Research and creator of Mathematica, WolframAlpha and the Wolfram Language, and Scott Worland, CTO at Norm AI.
Wolfram began with a history lesson. In the 1600s, Gottfried Leibniz, best known as a co-inventor of calculus, championed the philosophical language movement, an effort to represent everyday language in a mathematical form. Leibniz was especially interested in anchoring legal decisions in formal processes. He recognised that contracts rely on a precise symbolic language, but lacked any mechanism to translate between that structure and ordinary speech, and the project stalled.
Wolfram believes modern AI could now provide that missing translation layer. If AI can turn legal wording into symbolic form, he argued, it becomes possible to compute with it and extend reasoning well beyond human limits. This idea underpins his career-long mission to make the world computational. WolframAlpha, launched in 2009 when many had written AI off, converts natural language into symbolic language that can be calculated, and has since contributed to the development of major large language models.
However, Norm AI’s account of the session highlights a second challenge. Beyond the layer linking humans to computation, people also need a layer that explains how AI systems actually reach their conclusions. Wolfram pointed to mathematics, where users often cannot describe how an AI solved a theorem. An intermediate representation that is both precise and human-readable is what he is now working to build.
That need is most acute in complex areas where judgment matters. Currently, the only way to be certain what a system will do a trillion steps ahead is to run every one of those steps, because even systems with known rules can behave unexpectedly. Organisations are therefore left choosing between restricting AI to simple tasks or accepting the possibility of surprises.
Wolfram noted that society has long relied on unpredictable tools, from horses onwards, and that history consistently bends towards automation. Rather than displacing people, he argued, automation tends to open up new forms of work, as the mechanisation of US agriculture demonstrated. The legal sector, in his view, should embrace the same shift.
The conversation closed on agent-to-agent interaction. Wolfram suggested AI agents behave much like humans, albeit with greater efficiency and a wider range of options considered. This raises pressing questions for regulators and the legal industry: what language should laws for agents be written in, do agents have freedom of thought, and should rules target the actuation layer rather than internal reasoning? Answering these, Norm AI’s write-up concludes, is key to fully harnessing AI as the bridge between people and computation.
For more, read the full story here.
Copyright © 2026 FinTech Global
Copyright © 2026 RegTech Analyst
