AI Legal Ethics: New Guidance from the ABA
August 1, 2024
Privacy Plus+
Privacy, Technology and Perspective
On Monday of this week, the American Bar Association (ABA) issued its first formal opinion on the ethics of lawyers’ use of generative artificial intelligence (GAI).
ABA Formal Opinion 512 presents a 15-page, heavily footnoted, analysis of ethical considerations for attorneys using generative GAI tools in their legal practices. Here is a short, abbreviated summary:
· Competence: Attorneys must stay abreast of technological advancements and their implications for legal practice under ABA Model Rule 1.1. The Opinion advises lawyers to understand how GAI tools operate and the associated risks to ensure competent representation. The standard of competence required is a “reasonable understanding” of the GAI tool’s capabilities and limitations, without “uncritical reliance” on content created by it. The Opinion strikes a cautious tone, warning counsel that GAI tools are “only as good as their data and related infrastructure” and may produce “unreliable, incomplete, or discriminatory results,” may “combine otherwise accurate information in unexpected ways to yield false or inaccurate results,” and may be “prone to hallucinations.”
· Confidentiality: Lawyers have a fundamental duty to protect client confidentiality under Model Rule 1.6. Entering confidential client information into a GAI tool increases the risk that others using or receiving output from the tool may learn it. More interestingly, the Opinion observes that this may also be true within a firm. Even if a firm creates its own, proprietary GAI tool for its exclusive use, its lawyers should realize that if one of them inputs client-confidential information, others in the firm may inadvertently use it and even disclose it, thus defeating client expectations and ethical walls that may be in place.
The Opinion therefore urges obtaining clients’ “informed consent” before use of client-confidential matter in GAI tools. In the Opinion’s view, including in engagement letters a boilerplate consent to use GAI tools is not sufficient (though see Our Thoughts below) – lawyers should read the Terms of Use, privacy policies and other contractual terms for use of the GAI tools, and if necessary, confer with IT or cyber security experts.
· Communication: As expected, lawyers should inform clients about the use of GAI tools in their legal matters. Clients must be aware of how their information might be handled and the risks involved. Attorneys should obtain informed consent from clients regarding the use of AI, ensuring that clients understand and accept the potential risks. The Opinion suggests that client consultation is required to the extent “reasonably necessary” to permit a client to make informed decisions – which is heavily fact-dependent. Consultation would be needed, for example, when a lawyer proposes to input client information into a GAI tool, or to explain her fees, or when a GAI tool’s output will affect a significant decision in the representation (e.g. jury selection).
· Claims and Candor to the Court: Overlapping with the duty of competence, the Opinion expects lawyers to “review for accuracy all GAI outputs.”
· Supervision: Lawyers are responsible for supervising and controlling the use of GAI tools by their staff and within their firms. This includes ensuring that GAI tools are used appropriately and that data handling practices align with ethical and legal requirements. Lawyers must establish policies and training, and make “reasonable efforts” to ensure that they are followed. Interestingly, the Opinion extends this as a recommendation for all vendor management, recommending that before engaging a GAI vendor, lawyers should investigate the tool’s reliability, security measures (including limitations on the vendor’s liability), whether data will be kept after the engagement, and any claims of proprietary rights in it; and ensure that the tool is configured to protect confidentiality, that that commitment is enforceable, and that the lawyer will be notified in event of a breach.
· Fees: The Opinion applies familiar rules to GAI tools. For example, a lawyer may not charge their clients for time spent learning a technology she will use for clients generally, but if a client specifically requests the use of a particular AI tool in a matter, she may charge for learning that particular tool. Similarly, a lawyer may consider the cost of GAI to be part of her office overhead, or may charge for a portion of an expensive and proprietary GAI tool, or on a per-use basis if appropriate – all as fully explained and consented.
You can read Formal Opinion 505, dated July 29, 2024, by clicking on the following link:
Our Thoughts:
Increasing Variety of Guidance: We have previously covered AI guidance issued from California, Florida, and New York, and you can read those posts by clicking on the links that follow:
https://www.hoschmorris.com/privacy-plus-news/ai-legal-ethics
https://www.hoschmorris.com/privacy-plus-news/ai-legal-ethics-new-guidance-from-new-york
We are struck by the increasing variety of authorities weighing in on the proper use of GAI in the legal profession. Some approach GAI in different ways. For example, some courts require disclosure of their use in court filings, while others do not (e.g. the Fifth Circuit. See https://www.reuters.com/legal/transactional/5th-circuit-scraps-plans-adopt-ai-rule-after-lawyers-object-2024-06-10/); and some states (e.g. New York) require a general disclosure of GAI use in engagement letters), while others (e.g. West Virginia) say “boilerplate” is not enough.
But in general, all of the guidance seems to be trending in the same, cautious direction: be careful, give courts a heads-up, and get client consent. We suspect that a year or two from now, technical use and developments will overtake us, and most of 2024’s pronouncements – and the engagement letters and client advisories written based on them – will begin to look quaint and antiquated.
Impact on Lateral Hiring and Lawyer Mobility: We think the most interesting aspect of the ABA Opinion is its point that by using the same GAI tool, even a proprietary one, other lawyers in a firm may inadvertently use and disclose client-confidential information. The Opinion seems to treat this as a liability risk, which of course it is, but we see it also as one more complication to lateral hiring and lawyer mobility, especially in the larger firms. As more firms have tried to grow by lateral hiring, we believe the pressure to make lateral hiring easier – especially by reducing or eliminating the “imputed knowledge” rule of confidentiality, which holds that knowledge possessed by one lawyer in a firm is presumed to be known by all – has been building over the past few decades. We expect GAI use to add to this pressure, perhaps in unforeseeable ways.
Looking Forward: As we advance into an era dominated by AI, we hope the legal profession turns its full attention to the potential displacement of legal professionals and the broader implications for the practice of law. As AI tools become more sophisticated, there is a real risk that routine tasks previously performed by junior lawyers or paralegals will be completely automated, leading to job losses and reduced opportunities for career advancement within the profession. Further, AI’s ability to handle large volumes of data and perform precise tasks raises questions about the role of human judgment in legal practice, and the fiduciary nature of legal services. The legal profession must ensure that AI tools are used to enhance rather than replace human judgment, preserving the quality of legal representation and maintaining the ethical obligations that underpin the practice.
Hollywood actors and writers have negotiated protections for their work through their unions, setting limits on the use of AI in their fields. Similarly, the legal profession must consider strategies to protect against job displacement and ensure that advancements in technology do not come at the expense of its workforce, and the justice system more broadly.
---
Hosch & Morris, PLLC is a boutique law firm dedicated to data privacy and protection, cybersecurity, the Internet, and technology. Open the Future℠.