Code and Rules of Evidence Committee Advocates for Statutory Definition of Generative AI

AI Ethics Opinion Offers Insights, Yet Leaves Key Queries Unanswered Regarding Evidence Obligations

In response to a growing need to protect “the foundational principles of our legal system,” the Code and Rules of Evidence Committee is taking proactive steps by proposing a statutory definition for artificial intelligence, specifically generative AI.

During a virtual meeting held on October 17, the committee, comprised of legal experts, voted overwhelmingly with a tally of 28 to 1 in favor of the proposal. Following the vote, committee member Kristina Feher emphasized to the Standing Committee on Professionalism that the proposed definition serves as a “starting position” for addressing the complexities associated with AI.

Feher, who leads an AI working group, has exciting plans to produce a Continuing legal Education (CLE) program focused on this emerging technology slated to occur later this spring. She highlighted that many legal professionals in Florida are still in the early stages of becoming familiar with AI technology, stating, “A lot of people have a somewhat rough knowledge of what it is.”

The proposed statutory definition, which will be incorporated into §90.951 (5), delineates “Generative Artificial Intelligence” as evidence produced by a machine-based system capable of mimicking the structure and characteristics of input data based on a specific set of human-defined objectives. This encompasses the generation of various forms of synthetic content, including but not limited to writings, photographs, recordings, images, videos, audio, text, and other digital materials.

The Florida Bar is currently working to prepare the proposed definition for public comment and to distribute it among interested sections and committees for further examination and feedback.

In a significant development in January 2023, Florida became the first state to enact comprehensive ethical guidelines aimed at the responsible use of artificial intelligence within the legal profession. Notably, Ethics Opinion 24-1 advises attorneys to secure “the affected client’s informed consent prior to utilizing a third-party generative AI program when such use involves the disclosure of any confidential information.” This ethical framework is part of an ongoing effort to ensure that AI’s integration into legal practice does not undermine client rights or legal integrity.

The formation of this AI committee was prompted by the need for a clear ethical standard surrounding AI’s role in law, alongside proposed amendments to various Bar Rules including Chapter 4 regarding a lawyer’s Responsibilities and Rule 4-1.6 on Confidentiality. The proposals received the Supreme Court’s approval on August 29, as documented in In Re: Amendments to Rules Regulating The Florida Bar – Chapter 4, Case No. SC2024-0032.

Contrary to Bar rules, ethics opinions lack binding authority; however, comments made on Bar rules are considered explanatory and informative. Despite the usefulness of the AI-related ethics opinion, members of the Code and Rules of Evidence Committee highlighted that it “does not address obligations surrounding evidence and AI-generated evidence,” emphasizing the gaps that still exist in the legal framework.

Chair of the committee, Katelyn Knaak Johnston, tasked her subcommittee with three principal objectives: to establish a comprehensive definition of generative AI, to identify rules requiring amendments to manage the implications of discovery and notice when using generative AI, and to compile a list of professionalism issues that arise due to the advent of AI in legal practice.

The committee further acknowledged recent legislative actions in Florida, particularly with the adoption of section 106.145, Florida Statutes, which defines generative AI in the context of elections. This definition, effective from July 1, closely aligns with the proposed language of the Code and Rules of Evidence Committee, although it refrains from referencing “evidence.”

Moreover, the newly enacted elections statute mandates that any political advertisement, electioneering communication, or miscellaneous political advertisement, wholly or partially created using generative AI, must carry a disclaimer, shedding light on the origin of its content.

In a notable instance last November, Ocala-based attorney Gordon Glover, who is also a board member of the Florida Bar, reached out to the Code and Rules of Evidence Committee to advocate for the proposed definition. Glover pointed out that Florida is taking a leadership role in addressing AI-related concerns and stressed the importance of preserving the reliability upon which the legal system is founded. He notably raised concerns regarding the threat posed by AI-generated “deep fakes” to judicial integrity.

“Hyper-realistic digital manipulations of audio and video can create dangerously convincing yet entirely fabricated pieces of evidence that disrupt the reliability required in legal proceedings,” Glover cautioned.

In February, the Florida Conference of District Court of Appeal Judges voiced their concerns regarding AI’s implications for legal filings in a letter addressed to both the Rules of General Practice and Judicial Administration Committee and the Appellate Court Rules Committee. They stressed the necessity for amendments to Rule of General Practice and Judicial Administration 2.515 (Signature and Certificates of Attorneys and Parties) and Rule of Appellate Procedure 9.210 (Briefs) to ensure lawyers appropriately manage AI usage in court documents.

Fourth District Court of Appeal Chief Judge Mark Klingensmith, who serves as the conference chair, expressed the growing unease of appellate judges in light of several Florida lawyers submitting court filings that relied on AI, which produced “hallucinations” or citations to fictitious judicial opinions. This highlights the crucial need for a systematic approach to ensure accuracy and reliability in legal processes utilizing AI.

In urging the committee to act decisively, Glover stated, “The potential for AI to disrupt the legal evidentiary process is not merely theoretical; it is an imminent reality that requires immediate and thoughtful action.”

To further aid legal professionals in navigating the complexities associated with AI, the AI legalese decoder can play an instrumental role. By simplifying and clarifying legal language, the AI legalese decoder can assist attorneys in understanding AI-related regulations and evidence requirements more effectively. This tool can help demystify the implications of using AI in legal contexts, ensuring compliance with new definitions and ethical guidelines while maintaining the integrity and reliability of legal proceedings. By leveraging this innovative resource, attorneys can better equip themselves to address the challenges and opportunities that generative AI presents within the legal sector.