AI Legal Ethics: Guidance from California and Florida, and a Proposed Rule from the Fifth Circuit

February 22, 2024

Privacy Plus+

Privacy, Technology and Perspective

This week, let’s consider recent guidance and proposed rules from California, Florida, and the United States Fifth Circuit Court of Appeals, all addressing the ethical and practical implications of the use of artificial intelligence (AI) by lawyers.

California's Guidance on Generative AI

Recently, the State Bar of California's Standing Committee on Professional Responsibility and Conduct issued guidance for lawyers on the responsible use of Generative Artificial Intelligence in legal practice. This guidance offers broad principles on how to integrate AI responsibly into legal practice. Here's an overview:

• + Confidentiality and Security: Lawyers must ensure the generative AI products they use do not compromise client confidentiality or data security. Information input into AI can be used for other purposes. Lawyers should consider this risk when using such systems. Data should be anonymized to prevent identification of clients, and lawyers should consult with IT professionals or cybersecurity experts, and AI review systems’ terms of use / terms of service to safeguard the data.

• + Competence and Diligence: The guidance emphasizes the need for lawyers to be competent in their use of generative AI, recognizing the technology's limitations and potential for producing false, inaccurate, or biased outputs, necessitating competent and diligent review by lawyers. Lawyers must double-check outputs, and AI should not replace the lawyer's critical analysis and professional judgment.  Lawyers are advised to critically assess AI-generated outputs and ensure their accuracy and relevance to their clients' needs.

• + Legal Compliance: Lawyers using generative AI must comply with all relevant laws, including AI-specific laws, privacy laws, cross-border data transfer laws, intellectual property laws, and cybersecurity concerns. Lawyers must have a thorough understanding of how generative AI tools may impact their practice and their clients.

• +  Supervision: The guidance calls for managerial lawyers to establish clear policies on the use of generative AI within their firms, ensuring both lawyers and non-lawyers adhere to professional standards. Training on the ethical use of AI is also recommended.

• + Communication: Lawyers should be transparent with their clients about their intended use of generative AI, discussing the technology's benefits and risks as part of their duty to communicate effectively.

• + Billing Practices: While generative AI can improve efficiency, lawyers must bill ethically for work produced with its assistance, avoiding charges for time saved through AI use. The costs associated with generative AI may be passed to clients if done lawfully.

• + Accuracy and Candor: Before submitting any AI-generated work to the court, lawyers must verify its accuracy and ensure it meets all legal and ethical standards, correcting any errors or misleading information.

• + Addressing Bias: Given the potential for bias in AI training data, lawyers must be vigilant in recognizing and mitigating bias in AI-generated outputs, particularly in sensitive areas like client or employee screening.

•  + Cross-Jurisdictional Issues: Lawyers are reminded to consider the laws and regulations of all jurisdictions in which they are licensed, ensuring their use of AI complies with these diverse standards.

This guidance frames a comprehensive approach to integrating generative AI into legal practice, focusing on maintaining the high ethical standards expected of the profession while navigating the complexities introduced by this emerging technology.

A link to the California guidance follows:

https://www.calbar.ca.gov/Portals/0/documents/ethics/Generative-AI-Practical-Guidance.pdf

Florida's Opinion on Generative AI

The Florida Bar's Opinion 24-1, issued on January 19, 2024, complements the California guidance and also addresses specific aspects of generative AI use in legal practice, such as advertising and chatbot communication. Here are the key points that supplement those offered in the California guidance:

• + Lawyer Advertising and AI Chatbots – Disclaimer Required: The Florida opinion specifically addresses the use of generative AI chatbots in client communications and lawyer advertising. It mandates that chatbots include a disclaimer clarifying they are AI programs, not lawyers or law firm employees, to prevent misleading interactions and comply with advertising restrictions.

• + Oversight of Generative AI: The Florida opinion underscores the importance of oversight when using generative AI, drawing parallels to supervising nonlawyer assistants. Lawyers are responsible for ensuring that any generative AI-generated work product is accurate and ethical, like supervising a paralegal's work.

• + Use of AI – Informed Client Consent Recommended: The Florida opinion recommends obtaining the client's informed consent before using third-party generative AI programs that might involve disclosing confidential information.

• + Understanding AI Technology: Lawyers are advised to have a thorough understanding of generative AI technology, especially regarding whether the AI is "self-learning," which could raise additional confidentiality issues as the AI incorporates new inputs into its data.

• + Ethical Use in Billing Practices: The Florida opinion provides detailed guidance on billing practices related to generative AI use, emphasizing that lawyers must not engage in double-billing or charge for time saved through AI efficiency. It suggests considering alternative billing arrangements that reflect the cost savings and efficiencies gained through AI use.

A link to the Florida opinion follows: https://www.floridabar.org/etopinions/opinion-24-1/

The Fifth Circuit's Proposed AI Restrictions

The Fifth Circuit Court of Appeals has proposed a rule that would require lawyers to certify that they either did not use generative AI programs like ChatGPT to draft filings or that human beings reviewed AI-generated material for accuracy before filing. This proposal aims to ensure transparency and accountability in the use of AI in legal work, addressing risks such as "hallucinations," or incorrect outputs, ethical and regulatory concerns, and algorithm bias​​.

A link to the proposed rule follows: https://fingfx.thomsonreuters.com/gfx/legaldocs/mopajaxmava/11222023ai_5th.pdf

Our Thoughts

The new guidance and proposed rules reflect an important recognition of the need to integrate AI into legal practice responsibly. They underscore the indispensable role of human judgment in the legal profession, a quality that AI cannot replicate—at least, not currently.

Still, there are several points in the guidance and proposed rules that may need further consideration. For example, we are skeptical that confidentiality may be preserved by “anonymizing” data. There are many reasons why, and for a discussion, you can review our 2021 post, entitled “When Anonymous Isn’t” by clicking on the following link:

https://www.hoschmorris.com/privacy-plus-news/when-anonymous-isnt

We also wonder whether the additional disclosure requirement proposed by the Fifth Circuit is needed. Under Federal Rule of Civil Procedure 11, lawyers already represent that “the claims, defenses, and other legal contentions are warranted by existing law or by a nonfrivolous argument for extending, modifying, or reversing existing law or for establishing new law.” Therefore, it seems to us that citing AI-hallucinated cases would violate this Rule. Perhaps, however, it’s helpful to remind lawyers that this requirement is not alleviated when using AI?

Regardless we expect to see more guidance issued as issues associated with the use of generative AI in legal practice continue to emerge.

--- 

Hosch & Morris, PLLC is a boutique law firm dedicated to data privacy and protection, cybersecurity, the Internet, and technology. Open the Future℠.