In the realm of legal technology, the recent United States v. Heppner case has sparked a heated debate about the boundaries of attorney-client privilege and the work product doctrine in the age of AI. The case centers around Bradley Heppner, a financial services executive charged with securities fraud, who used an AI translator, Claude, to prepare for his defense. The question at hand: Does a defendant who uses an AI translator retain attorney-client privilege? The answer, according to Judge Jed Rakoff's decision, is a resounding no. But is this the right approach? This article delves into the intricacies of the case, the judge's reasoning, and the potential implications for the accessibility of legal services and the resource divide in the legal profession.
The Judge's Decision
Judge Rakoff's ruling in United States v. Heppner was a groundbreaking moment, as it was the first of its kind to address the issue of AI and privilege. The case arose when Heppner, after receiving a grand jury subpoena and retaining counsel, used Claude to prepare for his defense. The AI tool created reports that outlined defense strategies and helped him develop arguments based on the facts and the law. When the FBI raided his home and seized the documents, Heppner's lawyers tried to reclaim them, asserting attorney-client privilege and the work product doctrine. However, Judge Rakoff disagreed, citing two related legal doctrines.
The attorney-client privilege protects confidential communications between a client and a lawyer made for the purpose of obtaining legal advice. Judge Rakoff correctly concluded that there was no attorney-client relationship because Claude expressly disclaims providing legal advice and the existence of any fiduciary obligation. Similarly, the work product doctrine safeguards materials prepared in anticipation of litigation, typically by or at the direction of counsel. Since Heppner's counsel did not direct him to use Claude, the documents did not reflect counsel's legal strategy at the time they were created, and thus, were not protected under this doctrine.
Where the Judge Went Wrong
While Judge Rakoff's conclusion was correct, his reasoning was problematic and went beyond what was necessary to resolve the case. The judge's analysis of a defendant's right to prepare a defense rested on a company's terms of service, which is a significant oversight. This approach has implications for how AI tools influence the accessibility of legal services.
The judge's decision hinged on the fact that Anthropic's privacy policy permits data collection, training, and disclosure to governmental regulatory authorities. This led to the conclusion that Heppner did not have a reasonable expectation of confidentiality, thus waiving privilege when he shared information with Claude. However, this approach elevates formal terms over the spirit of the law of privilege and ignores the existing framework for evaluating confidentiality.
Courts have a well-established process for assessing whether sharing documents with a third-party service eliminates confidentiality. For instance, when a law firm or company stores privileged documents in the cloud, as long as the vendor takes reasonable steps to maintain confidentiality, the use of that platform does not waive privilege. The American Bar Association (ABA) has also provided guidance on this matter, allowing lawyers to transmit client information electronically as long as they take reasonable steps to safeguard it.
The Broader Implications
The consequences of Judge Rakoff's decision extend far beyond the Heppner case. As clients increasingly use AI to engage with legal matters, the potential for losing attorney-client privilege becomes a significant concern. Non-English-speaking defendants might use AI translators to communicate with counsel, or clients might use AI to organize and summarize records before a meeting with their lawyer. After the meeting, they might use AI to test arguments and prepare questions for the next session.
While some of these uses might be protected by the work product doctrine if directed by counsel, the broader range of situations where clients consult lawyers is not covered. Attorney-client privilege is the only protection in these cases, and Judge Rakoff's analysis undermines it. This exacerbates the resource divide, as those who can afford human intermediaries retain privilege, while those who cannot risk losing it.
A Way Forward
To address these concerns, future courts should focus on the particular facts of each case, including whether the materials were prepared for counsel, whether they reveal privileged communications, and whether the defendant took reasonable steps to maintain confidentiality. This approach aligns with existing doctrinal frameworks and avoids the complexities of AI-specific questions.
Furthermore, the ABA and other bar associations should issue new guidance on whether and how attorneys can direct clients to use AI tools without forfeiting privilege. Lawyers should be obligated to advise clients on using these tools, ensuring that confidentiality is maintained. Allowing the confidentiality question to be resolved case by case through terms-of-service analysis is a poor way to develop coherent policy.
In conclusion, the novelty of AI should not produce a result where the scope of a client's right to communicate with counsel and prepare a defense in confidence turns on their ability to afford a human intermediary. Courts must recognize the potential pitfalls of AI and take a balanced approach to protect both confidential information and the benefits of these tools.
