Law Firms Navigate AI Usage Disclosure and Client Consent Dilemmas

A new debate has emerged for law firms as generative artificial intelligence (AI) grows increasingly prevalent: to disclose or not to disclose their use of AI technology to clients. Firms like Cleary Gottlieb Steen & Hamilton are grappling with the question, though its managing partner, Michael Gerstenzang, insists full disclosure is key.

Furthermore, this question necessitates grappling with whether the use of certain AI tools needs to be disclosed. For example, does the law firm need to disclose its use of AI in their research tools and data processing procedures to their clients?

Different perspectives are emerging from legal bodies as well. The California Bar and Florida Bar recently provided guidance recommending lawyers consider disclosing AI’s use and obtain client consent respectively, particularly when sensitive data is involved. Furthermore, a Massachusetts Institute of Technology task force suggested that client engagement terms should address the responsible use of AI and debated whether the client’s consent should be required.

Katherine Forrest, a partner at Paul Weiss, suggested that lawyers should use discretion while dealing with confidential information and generative AI, especially in the light of potential leaks if the data is input into open, public AI systems.

The integration of generative AI into widely used programs, such as Microsoft Corp.’s AI Copilot, slowly blurs the line between regular technology and AI tools. According to Jeffrey Saviano, a member of the MIT task force, the disclosure question will become obsolete as AI becomes more widespread and integrated.

Firms are looking towards their clients for guidance around disclosing AI usage. Disclosure expectations may appear in engagement letters, playing an increasingly pivotal role in conversations about AI use. Notably, in-house lawyers have been asking firms about their use of AI technology, specifically with regard to increasing efficiency.

Finally, the responsibility hangs on the lawyers to follow their client’s instructions pertaining to AI usage as where and how it should be disclosed. Certain clients have asked the team at Paul Weiss not to use generative AI without communicating with them, for instance.

The full details of the debate can be found in the original article.