Benchmarks

Citation-accuracy and reliability benchmarks for AI legal research and drafting tools, framed as risk assessment rather than product endorsement. Each entry discloses the benchmark source type (peer-reviewed academic study, vendor-published report, or independent test), the tool name and version tested, the test date, and a methodology summary, and visually separates independent findings from vendor-coordinated ones. Entries cross-link to any Risk Digest case naming the same tool. Does not contain case outcomes (Risk Digest) or how-to guidance (Workflows), though it links to both.

ToolVendorSourceHallucination rateVersion testedTest date
Cory Mills investigations test four separate legal frameworks
Rape and Trafficking Complaints Against Daniel Siad Spur Two-Nation Probe
The Defense-Side Legal AI Opportunity: Why the Next $1B Startup May Come From an Underserved Market
A civil rights legal analysis of the Don Lemon arrest
FBI World Cup security legal response faces structural gaps
Federal Court AI Disclosure Rules: Which Tracker to Use and When
Comparing Flanagan and Craig on election law
Richard and Heather Altman Face Fraud Charges Across Five States
Free AI Tools for Lawyers: What Works, What Doesn't, and Where the Risks Are
How Courts Rejected the Policies Selena Gomez Protested
What the Hyundai Kona Recall Teaches About Product Liability
Instagram DMs Can Now Be Used as Court Evidence After Encryption Drop
What IREN's Stock Decline Tells Us About AI Data Center Investment
Three Legal Complexities in the Landen Moore Drowning Murder Charge
What the Surveys Show About Legal AI in 2026
The Legal Problems Behind Crocodile and Alligator Prisons
Le'Veon Bell's $25M Default Judgment Shows Service-of-Process Risks
Melania Trump Epstein denial: Wolff anti-SLAPP dismissal explained
NBA press conference leaks create dual legal exposure
How NDAA Section 219 Creates a New Executive Agent for Israel
Blogarama - Blog Directory