Writing on verification, vendor and architecture risk, and what has to be tested and documented before an AI system is relied on. Much of the current archive was written for a litigation audience and uses legal-sector examples.
Written for the people who approve, build and have to live with AI systems.
Analysis grounded in enterprise engineering and regulated-system experience.
Published by Rule26 AI · Research · Guidance · Analysis
Blog posts are for information and education, not legal or insurance advice.
Probabilistic AI is flexible but never guaranteed correct. A deterministic spine adds the verification, rule enforcement, and auditability that legal and safety-critical workflows demand.
Read article →Latest Research
Probabilistic AI is flexible but never guaranteed correct. A deterministic spine adds the verification, rule enforcement, and auditability that legal and safety-critical workflows demand.
Read article →An honest comparison of citation verification tools and methods, including tradeoffs, costs, and reliability for different practice areas.
Read article →Two free AI tools for law firms just launched. The efficiency gains are real. So is the risk. Here is what small firms need to know before touching them.
Read article →Why lawyers must read everything, and why AI checking AI is a dangerous illusion.
Read article →AI is already part of many lawyers' drafting workflows. It also confidently invents cases that look real. Here's a simple extra step to verify citations and quotes against the record, free, no AI required.
Read article →Traditional litigation wasn't built for the age of algorithms. Discover how the new JAMS AI Dispute Rules protect trade secrets and streamline discovery for litigators and tech companies alike.
Read article →Explore how The New York Times built an 'ironclad' evidentiary fortress against OpenAI and Microsoft, providing a masterclass for litigators and a wake-up call for the open web.
Read article →Publishing AI-generated content without meaningful review is not just a quality issue. It is a reputational and malpractice risk. A practical risk framework for lawyers and professionals who put their name on the line.
Read article →How the forensic tools used to catch pirates paved the way for a landmark AI music partnership.
Read article →The same questions run through every engagement: what the architecture assumes, what a vendor is really committing to, what can be independently verified, and what evidence is left behind. Book an initial technical review to work through yours.
Discuss Your Project