Your client trusted you with their secrets. Did you trust an AI with them too?

A bankruptcy attorney fought to keep a client's records out of discovery, arguing attorney-client privilege. The case he cited to make that argument didn't say what he claimed — it didn't even involve privilege.

In In re Rosslyn2016, LLC (S.D. Texas Bankruptcy Court, decided July 14, 2026), the attorney's discovery motion leaned on Westlaw Precision for its citations. The one meant to anchor the privilege argument, "In re Grand Jury Subpoena, 870 F.2d 343," turned out to be misused: the actual case at that reporter cite, Grubbs v. Norris, doesn't support the privilege proposition asserted. Several other quoted passages weren't in the opinions they were pulled from.

The court's response: civil contempt, mandatory CLE on generative AI, and an adverse costs order of $29,877.

Here's the part worth sitting with. The client's privilege was the exact thing this filing was supposed to protect. The tool used to build that protection never verified what it was citing.

Attorney-client privilege is the foundation of the relationship. When confidential strategy gets routed through an AI research tool without anyone checking its output, the risk isn't just sanctions — it's the thing your client actually paid you to safeguard.

Three questions worth asking at your firm this week:

Which active matters have used AI-assisted research in the last 90 days? Has every citation in those filings been independently verified against the source? Is there a written protocol for AI use before privileged or dispositive filings go out?

If the answer to any of these is "I'm not sure," that's the gap to close first.

Full case: In re Rosslyn2016, LLC, S.D. Tex. Bankr., decided 7/14/26. Source: Charlotin AI Hallucination Database (damiencharlotin.com/hallucinations).

#LegalAI #AttorneyClientPrivilege #LawFirmRisk #LegalEthics