42 AI Setups Took The Same Legal Quiz—One Scored 8/100

A test by Lexassay’s operator scored 42 AI setups on 15 legal questions. A law-focused model scored 8/100; the author discloses a grading conflict.

Facts May Be Free, But Publishers’ Writing Has Rights: The AI Question

An EU court adviser says chatbot answers can infringe publishers’ rights in some cases, but his nonbinding opinion weighs against the publisher in a Gemini dispute.

Z.ai’s Legal AI Benchmark Score Looks Low. Its Ranking Tells Another Story

GLM-5.3-FlashX-Legal scored 19.17% on an all-criteria benchmark, placing seventh of 81 models while passing 92.5% of individual criteria.

Korean Court Precedents And AI: What 360 Judges Reported

A South Korean court survey of 360 judges found many had encountered nonexistent or distorted precedents and called for stronger rules.

Refusing AI At Work: A Humanist’s Case And Its Legal Limits

The University of Kentucky approved David Durman’s request not to personally use AI at work, a case shaped by federal religious-accommodation law.

How AI Reporting Software Fits Into Private Investigations

A proposal for AI-assisted PI reports targets time-consuming writing, with attorney review and time-saved measures as key validation tests.

What’s At Stake In The Allegations Over Claude’s Training Data?

A 2025 ruling found Anthropic’s book downloads unlawful but its training fair use. New author lawsuits remain unresolved; a $1.5 billion settlement covers past claims.

Grok AI Privacy Claims: What The SpaceXAI Class Action Alleges

A Top Class Actions headline says a class action accuses Grok of violating Illinois privacy law, but the available material lacks case details.

Contractor Vetting Starts With Asking About Safety

A proposed contractor-rating service would grade license, insurance and public-record signals, but its accuracy and homeowner demand remain untested.

Why Ironclad’s Fine Print Matters As OpenAI Trains Agents In Software

OpenAI’s Ironclad trial reports 55% of rubric criteria met by GPT-6 Astra. The test points to progress, but not deployment-ready contract work.