AI Queries Can Turn Into Legal Traps

A CFO might ask an AI about antitrust risks before a board meeting, thinking it’s just a quick check.
But the answer could become evidence against her company if regulators step in.

The Bradley Heppner Case

• Background: Former Texas finance chief Bradley Heppner was charged in 2023 with defrauding investors by moving $150 million between two hidden companies.
• AI Use: While under investigation, he used an AI tool to draft legal‑strategy reports and claimed those documents were privileged.
• Court Ruling: When the FBI seized his devices, the judge ruled that the AI‑generated files did not meet any of the three conditions for attorney‑client privilege:
  1. No lawyer was involved.
  2. No expectation of confidentiality existed.
  3. The AI itself cannot give legal advice.

Takeaway: Any time a business uses an AI chatbot for legal questions, the resulting text can be discovered in court. If a merger later triggers an antitrust lawsuit, the opposing side could request that AI output and use it against the company.

Warner v. Gilbarco

• Scenario: A plaintiff used ChatGPT to prepare an employment‑discrimination claim.
• Outcome: The judge protected those materials under the work‑product doctrine because they were created in anticipation of litigation and no lawyer was involved.
• Lesson: AI does not automatically erase legal protections; it depends on how the tool is used.

Legal Foundations

Courts are applying these timeless rules to AI‑generated content, not creating new AI laws.

Practical Guidance for Executives

1. Document Usage: Record who creates or reviews AI‑generated material.
2. Confidentiality Controls: Keep access tight; treat it like any other privileged document.
3. Legal Review: Have a lawyer draft or review the material whenever possible.

Adhering to these practices can make a big difference if the documents later appear in discovery.