-
What Happens When an Expert Witness Lets ChatGPT Write the Report?
A Harris County jury awarded $61.5 million after discovering that 3M’s expert witness used ChatGPT to compose up to 90% of his report. 350 pages of prompt logs became evidence.
-
Beyond “The AI Said So”: Why Governance Infrastructure Is the Missing Layer for Regulated Professionals
Courts documented 1,598 AI hallucination cases by mid-2026. Policy documents alone cannot defend against regulatory scrutiny. Governance infrastructure is the missing layer.
-
Anthropic Studied 37,657 Real Conversations and Found the Gap We Are Building For
Anthropic studied 37,657 real conversations and found legal is highest-stakes with no safety evaluation.
-
600 Vulnerabilities a Month: Why the AI Security Boom Is a Governance Problem
Microsoft patched 600+ vulnerabilities in one month. For regulated professionals, the real risk is not the flaw but the governance gap.
-
The $5,000 Fine That Was Supposed to Be the Lesson: Mata v Avianca
A $5,000 fine for six fake cases was treated as an isolated embarrassment. Three years and 2,000 documented cases later, it was the first entry.
-
The SRA Warning Notice on AI: What Every Firm Needs to Do Now
The SRA published a Warning Notice on 17 August 2026 confirming that law firms using AI without proper oversight face disciplinary action. 42 reports in 12 months.
-
When an Expert’s ChatGPT Prompts Became the Evidence: Watson Grinding v 3M
An expert asked ChatGPT to show 3M was 0% at fault. 350 pages of prompts became evidence. The jury awarded $61.5 million.
-
Hidden Instructions in Court Filings: The First Prompt Injection Sanctions
Courts in Connecticut and Brazil sanctioned hidden AI instructions in filings. A legal research platform was also compromised.
-
Both Sides Filed Fake Citations. The Judge Cancelled the Trial: Withers v City of Aberdeen
A federal judge cancelled the trial and suspended both lead attorneys for two years after both sides filed AI-fabricated citations.
-
The Court Just Ordered an Expert to Hand Over Her AI Prompts: CLF v Shell
A federal magistrate ruled that AI prompts used to prepare an expert report are discoverable methodology under Rule 26(b).