Source: DLA Piper

DLA Piper analyzes a critical ethical and legal dilemma: the risk of waiving attorney-client privilege when using public generative AI tools. The publication explains that inputting sensitive or privileged legal advice into a public AI platform may be legally construed as a disclosure to a third party. Under many jurisdictions, such a disclosure can lead to an automatic waiver of privilege, potentially making the communication discoverable by opposing counsel in litigation. This risk is heightened by the fact that many AI providers use user inputs to further train their models, effectively making the data part of the public domain.

To address these concerns, the article strongly advises legal professionals to avoid using public, consumer-facing AI for privileged work. Instead, law firms and in-house legal departments should only utilize enterprise-grade AI solutions that provide guaranteed data isolation and security. The piece also stresses the need for clear internal policies that prohibit the sharing of confidential client information with any tool that does not offer a robust "no-training" guarantee. As courts begin to rule on these issues, maintaining the sanctity of professional privilege requires a sophisticated understanding of how data flows through large language models.