Another alert about AI—we know. But this one is worth the quick read: how your organization uses AI today carries direct, and often overlooked, negative consequences as we move into an AI-enabled world.

Large Language Models, or “LLMs,” are the technology running behind the scenes when you ask ChatGPT to put your dog in a funny outfit. Increasingly, they have become a singular resource for how attorneys and their clients draft and analyze documents and communications. Used without adequate confidentiality guardrails, LLMs can expose a franchise system’s proprietary information and even compromise the attorney-client privilege in a future dispute.

Earlier this year, the United States District Court for the Southern District of New York, in United States v. Heppner, ordered a criminal defendant to produce documents he generated while communicating with Anthropic’s well-known “Claude” program about his case. 820 F.Supp.3d 292 (S.D.N.Y. 2026.) Those documents included the analysis and advice his own attorneys had prepared. The court reasoned that, even though the prompts may have originated with the defendant’s attorneys, the privilege was lost: LLM platforms are not attorneys, do not provide legal advice, and the platform’s own terms did not guarantee confidentiality. The court further held that Claude’s output was not shielded by the work-product doctrine because it was not “prepared by or at the behest of counsel.”

The Heppner ruling is a wake-up call. Before dropping counsel’s advice or work product into publicly available tools like ChatGPT and Claude, franchisors—and their franchisees—should think twice. Franchise systems need AI policies at both the corporate and franchisee levels to maintain compliance and visibility. AI is driving efficiency and output at a compounding pace, but without proper governance, the downside can quickly outweigh the gains—starting with the very real risk of handing your worst-case view of a dispute straight to opposing counsel.