AI and eDiscovery: EDRM Considerations on Professional Responsibilities

Editor’s Note: The Electronic Discovery Reference Model (EDRM) is a widely recognized framework that outlines the stages and processes involved in electronic discovery, which refers to the identification, preservation, collection, processing, review, analysis, and production of electronically stored information (ESI) in legal matters. EDRM provides a structured approach for organizations and legal professionals to effectively manage the electronic discovery process. EDRM was developed in 2005 by a group of legal professionals, industry experts, and technology vendors who recognized the need for a standardized methodology to address the challenges associated with the increasing volume and complexity of electronic data in litigation and investigations. The framework has since become a leading resource in the legal industry and has been adopted by numerous organizations worldwide. The new paper titled “Professional Responsibility Considerations in AI for eDiscovery: Competence, Confidentiality, Privacy, and Ownership” was published by the EDRM in June 2023 and is beneficial as it provides comprehensive guidance on the ethical and professional responsibilities of legal practitioners when using AI for eDiscovery. It addresses key issues such as competence, confidentiality, privacy, and ownership, helping practitioners navigate the challenges and responsibilities associated with the use of AI and related technologies.

Background Note: The new paper titled “Professional Responsibility Considerations in AI for eDiscovery: Competence, Confidentiality, Privacy, and Ownership” provides a comprehensive analysis of the ethical and legal considerations surrounding the use of artificial intelligence (AI) in the field of law, particularly in eDiscovery.

The paper begins by emphasizing the importance of competence in using AI tools. It suggests that legal professionals need to understand how these tools work and how to use them effectively to extract, analyze, and apply information from large data sets. This competence is crucial in leveraging AI tools to provide new insights, systematize processes, speed up resolution times, and reduce costs.

Shining a light on the issue of confidentiality, the paper surfaces ethical questions about using confidential client information in AI models, particularly when these models are used iteratively over a series of similar matters. The paper warns of the potential risks of sensitive client information being accessed by third parties through the AI models, highlighting the need for stringent measures to protect client confidentiality when using AI tools in legal practice.

In addressing the evolving landscape of privacy laws and regulations, the paper points to the need for attorneys to be aware of these regulations, especially when they apply to data consumed or produced by AI tools. The paper discusses privacy regulations’ potential dilemmas, particularly when an organization no longer holds the original data but related data lives on in a predictive model. This dilemma raises complex questions about data retention, deletion, and the potential privacy risks associated with retaining client data.

The paper also tackles the question of ownership, discussing who owns the economic value derived from AI models trained on client data. It raises questions about whether clients should have a stake in the value derived from these models and whether the models constitute attorney work product. This concern highlights the need for clear guidelines on the ownership and use of AI models in legal practice.

Emphasizing AI’s potential benefits in improving legal practice’s efficiency and effectiveness, the paper also warns of potential pitfalls as new regulatory frameworks emerge and case law provides precedents on the appropriate application of AI and related technologies. It advises attorneys to be aware of these shifting challenges and their attendant responsibilities.

With this paper, EDRM provides a reasoned guide for cybersecurity, information governance, and eDiscovery professionals, highlighting the ethical and legal considerations of using AI tools in legal practice. It accentuates the need for professionals in these fields to be competent in using these tools, to protect client confidentiality and privacy, and to understand the implications of AI on the ownership of economic value derived from client data.

Announcement and Paper*

Professional Responsibility Considerations in AI for eDiscovery: Competence, Confidentiality, Privacy, and Ownership

EDRM Project Team Publishes Final Paper After Public Comment Period

Setting the global standards for eDiscovery, the Electronic Discovery Reference Model (EDRM) is pleased to announce its white paper addressing “Professional Responsibility Considerations in AI for eDiscovery: Competence, Confidentiality, Privacy and Ownership is now final after a public comment period.

Privacy and security-enabled eDiscovery technology and services will play a central role in organizations as they strive to rebound from the current global crisis as more cohesive, more stable and more productive entities. Robust eDiscovery technology, supported by expertise and services, helps organizations respond to litigation, compliance and investigation demands. In a new era of remote work at scale, organizations need to securely manage information, protect confidential and privileged access and collaborate effectively to be efficient and productive and adhere to the increasing complex demands of security and data privacy frameworks.

Use of artificial intelligence (“AI”) tools in eDiscovery creates new opportunities for attorneys. By extracting, analyzing, and applying information from large data sets, AI tools can provide new insights, systematize processes, speed time to resolution, and reduce costs. A notable example is technology-assisted review (“TAR”), a process that makes use of machine learning to prioritize or classify relevant material in document reviews. Legal practitioners may reduce costs, time, and mistakes by applying TAR in litigation, antitrust reviews, investigations, and other matters. However, as legal teams’ uses of these technologies evolve, ethical issues may arise, particularly with the opportunities for reusing the results of the computer learning in future matters, but for different clients.

The white paper addressing “Professional Responsibility Considerations in AI for eDiscovery: Competence, Confidentiality, Privacy and Ownership” is published by EDRM’s Analytics and Machine Learning’s subgroup on AI Ethics and Bias, led by Project Trustees, Khrys McKinney, Principal, K L McKinney and Dave Lewis, Chief Scientific Officer, Redgrave Data.

“Attorneys who authorize the use of machine learning on their client data can improve their ability to protect themselves and their clients by first learning what, if any, of their client’s information will possibly inform algorithms beyond the initial matter,” said Khrys McKinney.

“AI programs like ChatGPT say the darndest things. So do machine learning systems that attorneys might train on client data, and it behooves them to be aware of the risks to confidentiality, privacy, and intellectual property,” asserted Dave Lewis. “We hope this white paper will provide helpful guidance.”

“EDRM and our whole legal community are fortunate that top eDiscovery data scientists, and other experts, led and contributed to preparation of this cutting-edge paper,” commented David R. Cohen, Reed Smith partner and Chair of the EDRM Project Trustees. “It highlights important issues that arise from the use of AI models across multiple matters– issues that most attorneys may not have previously considered, but must consider– to ensure that we are fulfilling our ethical duties, including protecting client confidences.”

EDRM is grateful to the project team (organizations noted for identification purposes only):

• Ricardo Baeza-Yates, Director of Research, Institute for Experiential AI at Northeastern University, USA (San Jose, CA)
• Lilith Bat-Leah, Vice President, Data Services at Digital Prism Advisors, Inc. (New York, NY)
• Darius Bennett, Darius Emeka Bennett, P.C., CEO and Attorney, Civil Litigation, eDiscovery and Criminal Defense (Birmingham, AL)
• Tara Emory, Senior Vice President of Strategic Growth and General Counsel at Redgrave Data (Falls Church, VA)
• David D. Lewis, Chief Scientific Officer at Redgrave Data (Denver, CO) [Trustee]
• Khrhysna McKinney, Principal at K L McKinney (Sugar Land, TX) [Trustee]
• Dana Bucy Miller, Associate Director, Legal Solutions, QuisLex Inc. (Baltimore, MD)
• James A. Sherer, Partner, BakerHostetler (New York, NY)
• George Socha, Senior Vice President of Brand Awareness, Reveal (St Paul, MN)

Among the EDRM opportunities and resources available are the ability to connect, network and contribute via EDRM projects and events, share their expertise with our global community. The EDRM community of knowledgeable, multidisciplinary professionals is building resources to enhance e-discovery, privacy, security and information governance frameworks, processes and standards.

The EDRM community is comprised of 33% corporations, 30% law firms and 23% software and service providers, 12% governments with the remaining 2% being a mix of educators, students, judges and media in 145 countries spanning six continents.

About EDRM

Empowering the global leaders of eDiscovery, the Electronic Discovery Reference Model (EDRM) creates practical resources to improve eDiscovery, privacy, security and information governance. Since 2005, EDRM has delivered leadership, standards, tools and guides to improve best practices throughout the world. EDRM has an international presence in 145 countries spanning six continents and growing has an innovative support infrastructure for individuals, law firms, corporations and government organizations seeking to improve the practice and provision of data and legal discovery. Learn more about the EDRM today at EDRM.net.

About EDRM’s Analytics and Machine Learning Project’s subgroup on AI Ethics and Bias

Use of artificial intelligence (“AI”) tools in eDiscovery creates new opportunities for attorneys. By extracting, analyzing, and applying information from large data sets, AI tools can provide new insights, systematize processes, speed time to resolution, and reduce costs. A notable example is technology-assisted review (“TAR”), a process that makes use of machine learning to prioritize or classify relevant material in document reviews. Legal practitioners may reduce costs, time, and mistakes by applying TAR in litigation, antitrust reviews, investigations, and other matters. However, as legal teams’ uses of these technologies evolve, ethical issues may arise, particularly with the opportunities for reusing the results of the computer learning in future matters, but for different clients.

Read the original announcement.

Complete Report: Professional Responsibility Considerations in AI for eDiscovery: Competence, Confidentiality, Privacy, and Ownership – EDRM (PDF) – Mouseover to Scroll

Read the original paper.

*Shared with permission from EDRM.net under Creative Commons – Attribution 4.0 International (CC BY 4.0) – license.

Assisted by GAI and LLM Technologies

Additional Reading

• Good Cybersecurity Practices for AI? A Multilayer Framework (ENISA)
• Exploring the Inclusion of eDiscovery-Centric Resources in the Google C4 Dataset: A Highly Selective Search

Source: ComplexDiscovery