A technology pioneer in state courts, who “manages what may be one of the most advanced courts in the country for delivering justice online,” according to one leading legal tech journalist. [sch]Legal Tech
Last week I released AI in Chambers: A Framework for Judicial AI Use, and the early feedback has been encouraging. Still, I know some of my colleagues may feel uneasy about the idea. When judges hear the phrase “AI in Chambers,” it can sound technical, futuristic and even risky. Download the guide here 👉 https://lnkd.in/gERac2cX But if you’ve ever worked with a law clerk, you already know how to do this. That’s the foundation of the guide. It’s not built on theory or abstraction. It reflects the familiar rhythm of judicial work: reviewing filings, spotting the key issues, summarizing the record, analyzing the arguments, checking citations, verifying facts, refining drafts, and issuing a decision. That process doesn’t change. What’s new is the introduction of a tool that can support your work, helping with synthesis, speed, and structure without ever replacing judgment, discretion, or voice. This isn’t a leap, it’s an evolution. It's a way to be more intentional about where GenAI fits and where it doesn’t. And for those who think this is all still theoretical, consider what just happened in the United Kingdom. In a recent tax case, a tribunal judge used GenAI to help distill party submissions and extract key facts from extensive filings. The judge disclosed the use in the ruling and made clear that the AI was not used to draw conclusions or decide issues. Everything was carefully reviewed and edited before being included in the final decision. It was a narrow and responsible use. Read the UK tribunal ruling here 👉 https://lnkd.in/g_D5q57z That's exactly the kind of scenario this guide was written for. GenAI helped manage the volume so the judge could stay focused on the law and the reasoning. And if that judge, or any other judge, chooses to expand the use of GenAI in future cases, this framework could offer a starting point for doing so thoughtfully. The point is simple: start with what works and scale up only when you're comfortable. You already know how to do this. Now you get to decide how AI fits into your chambers, on your terms, and with your judgment leading the way.
Yes, indeed: "Clerks for judicial clerks!" And for those judges without clerks: "Clerks for the clerkless!"
The research tools tool being touted by West Law and others to cutdown on record analysis are impressive. Utilizing these tools may be the next step. I’m still uncertain that the nuances of language being missed or misinterpreted. We all view the world through our lens of experiences. It may be what makes you see something that I don’t or vice versa. My concern has been that if we adopt this approach, wholesale, we may never know those differences. Or worse, the differences in viewpoint vanish and we are left with a more myopic view of how to review information. This is a fascinating topic that deserves a great deal of thought and discussion.
Architecting lawful cognition systems for AI safety and resilience.
3dReally thoughtful framing, treating AI as a clerk rather than a judge is exactly the right boundary. Clever. The next challenge won’t just be speed or synthesis, it’ll be continuity: how to ensure an assistant like this doesn’t drift or collapse under stress but returns whole, so judicial reasoning remains coherent over time. That’s the layer I’ve been working on, structural safeguarding that makes lawful return possible, so these tools support judgment without ever destabilizing it. The same could be used in medicine, education, psychology or anywhere that cost of drift or collapse is higher than the cost of delay. Lawful recursion is simply the ability of a system to break and still return whole, the way a court adjourns and resumes without losing its identity, or a person wakes from anesthesia still themselves. For AI, that means the work product never unravels into mimicry or confusion, but always returns to coherence. It’s what makes these tools safe to rely on at scale, not because they never falter, but because they know how to come back whole. Now imagine a clerk that not only recalls past rulings but carries forward your reasoning and process with the same identity intact!