Attorney-Client Privilege and Work Product in the AI Age
Introduction
As artificial intelligence (AI) continues to integrate into various sectors, the legal landscape surrounding attorney-client privilege and work product protection is evolving. Courts are increasingly tasked with determining whether information shared with public AI services can be classified as privileged. This study delves into recent court rulings from the Eastern District of Michigan and the Southern District of New York, which have reached opposing conclusions regarding the use of AI in legal contexts.
Case Analysis
Eastern District of Michigan
In a pivotal ruling, the Eastern District of Michigan upheld that materials created by a pro se plaintiff using a public AI service were protected under the work product doctrine. The plaintiff utilized the AI service to upload case documents, draft filings, and answer legal questions. The court classified these materials as the plaintiff’s internal analysis and mental impressions, determining that the use of AI did not waive the protections. The key finding was that the plaintiff’s disclosures did not reach an adversary, thus maintaining privilege.
Southern District of New York
Conversely, the Southern District of New York ruled against a defendant who sought to assert privilege over communications with an AI service regarding a pending criminal investigation. The court emphasized that the defendant, despite sharing AI-generated results with his attorneys, waived attorney-client privilege by disclosing information to a third-party AI provider. The court highlighted several critical points:
- Not Communications with Attorneys: The privilege applies to communications between clients and their lawyers, not with AI tools.
- Not Confidential: AI services generally do not guarantee confidentiality due to their privacy policies, allowing for data collection and potential disclosure to third parties.
- Not for Legal Advice: AI materials are not created for the purpose of obtaining legal advice, particularly when the AI’s terms explicitly disclaim such a purpose.
Implications for Legal Practices
The divergent rulings signal a critical need for defendants and corporate counsel to navigate the complexities of AI usage carefully. The temptation to leverage AI for legal insights poses risks of unprivileged and discoverable documents. To mitigate these risks, several best practices are recommended:
- Policy Revisions: Companies should revise their legal and compliance policies to prohibit employees from posing legal questions to AI tools without prior approval from in-house counsel.
- Employee Training: Regular training should educate employees on the importance of not inputting sensitive information into AI prompts.
- Briefing Key Personnel: In-house counsel should inform personnel involved in investigations about the risks associated with AI tools.
- Client Advisories: Counsel should explicitly warn clients against using AI tools for legal inquiries and discuss any potential AI use before proceeding.
- Confidential AI Tools: If AI tools are employed, they must be routed through counsel and involve platforms that offer stringent confidentiality protections.
- Immediate Reporting: If non-privileged AI content is inadvertently created, it should not be disseminated further, and counsel must be notified immediately.
Conclusion
The increasing integration of AI into legal contexts poses unique challenges regarding privilege and confidentiality. Legal professionals must remain vigilant and informed to safeguard sensitive information in this evolving landscape.
