Articles

Anthropic announced a three-layer guardrail stack to prevent hallucinated citations. When the bar opens a grievance, it will not ask what the guardrails did. It will ask what you did. Why guardrails and governance protect different things - and what a working governance program looks like for a small firm.

Governance, Not Guardrails

Anthropic's three-layer hallucination stack is impressive engineering. It is not a governance program. Why the bar regulates lawyers, not vendors - and the six artifacts a small firm needs on file before a grievance lands.

You Don't Ask the Liar If He's Lying

One hallucination, one bibliography, then the citation graph carries it forever. Anthropic just launched Claude for Legal with a plugin called AI Governance Legal. A vendor cannot independently audit its own product.

Four Factors Decide Whether an AI Tool Belongs in Your Firm. Most Reviews Cover Two.

Most AI tool reviews for law firms cover features and cost. They skip the two factors that decide whether the tool can touch a client matter. The Four-Factor Framework and the patterns it surfaces across 64 vendors.

The Same Legal AI Tool Can Be Both Prohibited and Authorized

Will Chen's open-source Mike OSS scores anywhere from Prohibited to Conditionally Authorized depending on how a firm deploys it. The four-factor framework on a deployment-dependent tool.

57 of 63: The Greg Lake Suspension and the End of the ‘Bad Lawyer’ Defense

Nebraska indefinitely suspended an attorney for AI hallucinations and concealment. Days later, Sullivan & Cromwell apologized for the same pattern. Eight questions every small firm needs to answer.

The Sullivan & Cromwell AI Filing: Where the Workflow Actually Broke

Three internal systems failed on the same filing. The five-layer workflow that closes the matter-level gap.

A Federal Judge Just Wrote Your Firm's AI Policy

The most comprehensive AI-in-litigation ruling to date. The protective order is a governance blueprint.

The AI Law Firm Debate Is Missing the Point

The real opportunity isn't building new AI law firms. It's building governance infrastructure for the 450,000 firms that already exist.

25 New AI Laws in 2026 - And Most Businesses Haven't Read Any of Them

Seven states, 25 new laws, 27 more pending. The state-by-state breakdown every business and law firm needs.

Florida Just Drew a Line on AI in Court

An appellate sanctions warning, mandatory disclosure orders, and a clear message: AI is permitted. Unchecked AI is not.

Same Question. Same Day. Opposite Answers.

Two federal courts split on whether AI-generated documents are privileged. The implications for every lawyer using AI.

The Hidden Risk in Legal AI Isn't Hallucinations - It's Informal Use

The real exposure for most firms isn't a fabricated citation. It's attorneys and staff using AI at will, with no policy and no record.

The Three Questions Every Law Firm Must Ask Before Using Any AI Tool

Before evaluating any product, three threshold questions - retention, supervision, workflow - must be answered first.

What Your Firm Owes Its Clients Before Using AI

Competence and confidentiality obligations do not pause for new technology. What lawyers owe clients when AI enters their practice.