Heppner & ABA Op. 512

What the case was. A criminal defendant used Claude, a consumer-facing generative-AI platform, to create materials related to his defense. He later sought to shield those communications and outputs from disclosure on attorney-client privilege and work-product grounds.

What the court held. Neither protection applied. The materials lacked confidentiality: the AI platform's terms permitted it to retain user inputs and outputs and reserved the right to disclose them to third parties, including government authorities. And the work-product doctrine did not reach them, because the materials were not created at the direction of counsel.

Privilege is not a default setting. In Heppner, neither was present.

What Heppner does not say. The ruling does not address how law firms should manage AI, what policies they should adopt, or what counts as "unmanaged" attorney use. It is narrowly a privilege case, decided on the specific facts of a defendant's own interaction with a consumer AI tool.

What firms can reasonably draw from it. When content touching a client's matter is fed into a consumer AI platform whose terms permit retention or disclosure to third parties, there is now federal court precedent that privilege is unlikely to attach to those materials. The cautious posture is to understand, for each AI tool in use, exactly what its terms permit — and to design workflows accordingly.

The American Bar Association's Standing Committee on Ethics and Professional Responsibility issued Formal Opinion 512 in July 2024. It is the first comprehensive guidance tying generative AI use to the Model Rules of Professional Conduct.

The opinion reads several familiar duties onto AI:

• ◆Competence (Rule 1.1). Lawyers must understand, at least at a reasonable-technology level, the capabilities and limitations of the AI tools they use.
• ◆Confidentiality (Rule 1.6). Inputting client information into a tool that retains, trains on, or exposes that data is a potential confidentiality breach. Informed client consent may be required.
• ◆Communication (Rules 1.2 & 1.4). Clients may be entitled to know that AI is being used on their matter.
• ◆Supervision (Rules 5.1 & 5.3). Firms must supervise how associates, paralegals, and vendors use AI — including "shadow" usage nobody formally approved.
• ◆Fees (Rule 1.5). Fees should reflect the time actually spent; efficiency gains from AI cannot be invisibly billed as attorney hours.

Op. 512 doesn't forbid AI. It requires lawyers to treat it like any other tool that touches privileged work: deliberately, with controls.