ediscovery Tag Archives — Kang Haggerty News

Legal Intelligencer: Metadata Attorneys Are Not Asking for (but Should Be)

Kelly Lavelle — Wed, 17 Jun 2026 19:14:04 +0000

Metadata can reveal information about a document’s origin, authorship, transmission, modification, and use that may not be apparent from the document itself. This broader understanding of metadata illustrates why limiting requests to a handful of standard load-file fields may overlook metadata that is far more significant to the claims and defenses at issue.

In the June 11, 2026 edition of The Legal Intelligencer, Kelly Lavelle writes, “Metadata Attorneys Are Not Asking for (but Should Be).”Most attorneys recognize the importance of metadata and routinely request standard fields such as creation dates, modification dates, authors and email transmission details. But those fields are not always the metadata that matters most. In a wrongful termination case, for example, counsel may request an employee’s personnel file and metadata showing when a disciplinary memo was created.

The document may be produced with those standard fields, yet the more important evidence could be the version history, which might show that the memo was substantially revised after the employee complained of discrimination.

Likewise, in a trade secret case, counsel may obtain emails and their metadata but fail to request audit logs showing who downloaded files before leaving the company. The strongest evidence may be in those logs, but it remains undiscovered because no one asked for it. The real issue is often not whether metadata was requested, but whether the right metadata was requested.

Part of the problem is that metadata is often discussed as though it consists solely of creation dates, modification dates, and author information. In reality, metadata encompasses a much broader range of information about a document’s history, characteristics and use. Electronic files may contain hundreds or even thousands of metadata fields. For example, email metadata can include information regarding transmission, receipt, forwarding, bcc recipients, and address-book data, while word-processing documents may contain edit histories, comments, formatting information, and other embedded data not visible on the face of the document.

As organizations increasingly conduct business through cloud-based collaboration platforms, messaging applications, and enterprise systems, some of the most significant metadata may reside in sources that rarely appear in standard discovery requests, including version histories, audit logs, access records, edit histories, and communication metadata.

Collaboration platforms such as Microsoft SharePoint, Google Workspace and similar systems illustrate this point. When a document is produced in litigation, counsel often receives only the final version of the document. What is frequently missing, however, is metadata showing how the document evolved over time.

Prior versions may reveal significant edits, identify additional custodians, establish when information was added or removed, and provide insight into the decision-making process. In cases involving contract negotiations, employment decisions, or internal investigations, version histories may tell a far different story than the final document ultimately produced.

Version histories can be particularly important because they may be lost when information is exported, converted, or produced in static formats. By the time discovery begins, the metadata that reveals how a document evolved may already be gone unless counsel identifies and preserves it early in the case.

Comments and tracked changes present a similar issue. Although attorneys routinely request final documents, they often fail to specifically address embedded comments, suggestions, and revision histories. These materials can identify individuals involved in drafting, reveal concerns later omitted from the final document, and provide contemporaneous evidence regarding a party’s knowledge or intent. Once documents are exported or converted for production, much of this information may be lost unless it is specifically preserved and requested.

Another frequently overlooked source of metadata is audit log data. Many organizations maintain logs showing when users accessed documents, modified files, downloaded information, or changed permissions. These logs can become particularly important when claims involve alleged document destruction, misuse of confidential information, trade secret disputes, or questions regarding the authenticity of evidence.

In some cases, audit logs provide the only reliable record showing who interacted with a document and when those interactions occurred. They may also identify individuals who never appear as authors, recipients, or custodians in the documents themselves but nevertheless played a significant role in the underlying events.

Communication platforms present a similar issue. Messages from Microsoft Teams, Slack, and similar applications are frequently produced as screenshots, PDFs, or other static images. While those productions may capture the visible content of a communication, they often omit metadata that may show whether a message was edited or deleted, identify all participants in a conversation, establish the precise timing of communications, or place individual messages within the context of a larger conversation thread. As a result, a screenshot may reveal what was ultimately said while hiding important information about the circumstances surrounding the communication.

Metadata can also reveal relationships that are not apparent from reviewing individual documents in isolation. Email and messaging metadata may identify communication patterns among key actors, reveal previously unknown participants in a discussion, and help reconstruct how information moved through an organization.

In some matters, understanding who communicated with whom, when, and how frequently may be as important as the content of the communications themselves. Such information can assist counsel in identifying additional custodians, evaluating witness credibility, and developing more targeted discovery strategies.

As organizations continue to rely upon cloud-based collaboration tools, the gap between the information that attorneys request and the information that actually exists will continue to grow. The Sedona Principles likewise recognize that the “special characteristics” of ESI, including metadata and other non-apparent information, may be pertinent to the form in which ESI should be preserved and produced. See “The Sedona Principles, Third Edition: Best Practices, Recommendations & Principles for Addressing Electronic Document Production,” Principle 12, Comment 12.a, 19 Sedona Conf. J. 169 (2018).

For that reason, counsel should consider metadata issues at the outset of a matter and identify the metadata most likely to prove or disprove the claims at issue, before relevant information is altered, lost, or produced in a form that no longer contains the information necessary to tell the complete story. In many cases, standard metadata fields will be sufficient. In others, version histories, audit logs, collaboration platform metadata, and document revision records may contain some of the most significant evidence in the case.

Kelly A. Lavelle is Senior Counsel at Kang Haggerty. She focuses on e-discovery and information management, from preservation and collection to review and production of large volumes of electronically stored information. Contact her at klavelle@kanghaggerty.com.

Reprinted with permission from the June 11, 2026 edition of “The Legal Intelligencer” © 2026 ALM Global, LLC. All rights reserved. Further duplication without permission is prohibited. Request academic re-use from www.copyright.com. All other uses, submit a request to asset-and-logo-licensing@alm.com. For more information visit Asset & Logo Licensing.

Legal Intelligencer: AI Systems and the Question of Confidentiality

Kelly Lavelle — Thu, 19 Mar 2026 14:44:18 +0000

Before privilege or production issues arise, the more basic inquiry is whether the information was ever confidential in the first place. This question is particularly significant in e-discovery, where electronically stored information generated by AI systems may later become discoverable.

In the March 19, 2026 edition of The Legal Intelligencer, Kelly Lavelle writes, “AI Systems and the Question of Confidentiality.”Much of the attention surrounding generative AI in litigation has focused on hallucinated authorities and AI-generated filings. But a different litigation risk arises when lawyers, or even clients themselves, enter client information into an AI platform. Before privilege or production issues arise, the more basic inquiry is whether the information was ever confidential in the first place. This question is particularly significant in e-discovery, where electronically stored information generated by AI systems may later become discoverable.

The U.S. District Court for the Southern District of New York’s decision in United States v. Heppner brings that issue into focus. Although the case has been described as a privilege ruling, the court’s analysis turned on a more basic issue: whether the information was confidential when it was entered into the AI platform. If client information is submitted to a system under terms that allow the provider to access, retain, or disclose it, the protections that depend on confidentiality, including attorney-client privilege and work product, may never attach.

In Heppner, a corporate executive under criminal investigation used the publicly available generative AI platform “Claude” to generate written analyses of potential defenses after receiving a grand jury subpoena. He later shared those documents with counsel. Following his arrest, FBI agents executed a search warrant at Bradley Heppner’s home and seized numerous documents and electronic devices. Heppner’s counsel later informed the government that among the seized materials were approximately thirty-one documents memorializing Heppner’s communications with the AI platform.

Through his counsel, Heppner asserted privilege over these documents, arguing that the information entered into the AI platform was learned, that he created the documents to facilitate discussions with counsel and obtain legal advice, and that he subsequently shared the AI outputs with his attorneys. His counsel acknowledged, however, that they did not direct Heppner to conduct the AI searches. The court rejected those claims. It held that the documents were not communications with an attorney, were not prepared by or at the direction of counsel, and, critically, were not confidential in light of the platform’s privacy policy.

The court’s analysis centered on the conditions under which the information was entered into the system. The provider’s terms stated that user inputs and outputs were collected and retained and that the company reserved the right to disclose data to third parties, including governmental authorities. Under those conditions, the court concluded that Heppner lacked a reasonable expectation of confidentiality at the moment of disclosure.

The court also rejected Heppner’s contention that he communicated with Claude for the “express purpose of talking to counsel.” In doing so, the court looked to the platform’s own representations. Claude expressly disclaimed providing legal advice. When asked whether it could provide such advice, the system responded that it was not a lawyer, could not offer formal legal recommendations, and advised users to consult a qualified attorney.

The court acknowledged that the implications of artificial intelligence for the law are only beginning to be explored. The court emphasized that AI’s novelty does not place it outside established legal principles, including those governing attorney-client privilege and the work-product doctrine. AI does not change the basic rules.

The opinion highlights a familiar but often underemphasized principle that privilege attaches only when communications are intended to be confidential and were, in fact, kept confidential. In the AI context, if a platform’s terms allow provider use beyond delivering the requested service, such as model training, analytics, affiliate sharing, or extended retention, the user may be acting inconsistently with the intent to preserve confidentiality at the moment of disclosure. The issue is not a subsequent waiver. It is that the platform itself may defeat confidentiality at inception.

This reasoning extends beyond generative AI. Courts have consistently held that employees may waive attorney-client privilege when using employer email systems to communicate with counsel, with the analysis turning on whether the employee had an objectively reasonable expectation of confidentiality. In those cases, the inquiry focuses on whether the intent to communicate in confidence was objectively reasonable under the circumstances. Heppner suggests that AI platforms may be subject to similar scrutiny. The terms governing data retention and disclosure may determine whether any protection exists at all.

The court also reaffirmed that privilege remains grounded in professional accountability. Recognized privileges depend on a relationship with a licensed professional who owes fiduciary duties and is subject to discipline. Generative AI does not occupy that role. Even if an output resembles legal advice, it is not a communication with counsel.

However, the court suggested that if counsel had directed the use of the AI tool within a structured framework, the analysis might have been different. That distinction highlights the importance of control. When AI use is integrated into a counsel-directed process under defined terms, protection arguments may be stronger. When it is independent and unsupervised, confidentiality may fail at the start.

Platform terms of service are therefore central. Many providers claim licenses to user inputs and outputs broader than a confidentiality-preserving relationship would tolerate, sometimes including rights to use content for service improvement, analytics or model training. However, this concern does not apply in the same way to all AI tools. Paid legal research platforms or firm-licensed systems are often governed by professional services agreements that include confidentiality obligations and defined data controls. In those contexts, the vendor functions more like a traditional litigation support provider than a public AI service. Even there, however, the analysis may turn on the contract. Courts will look at the terms to determine what they permit and how the provider handles user content.

These issues directly affect attorney-client privilege. Privilege protects confidential communications made for the purpose of obtaining or providing legal advice. The problem is not simply waiver. It is that the basic requirements for privilege may never have been satisfied. The work-product doctrine raises a similar issue. Work product protects materials prepared in anticipation of litigation. If those materials are created on a public nonconfidential platform, it weakens the argument that they reflect protected legal strategy shielded from disclosure.

Heppner reinforces the basic rule that privilege depends on confidentiality at the time the communication is made. It does not arise simply because a document is later shared with counsel. In the AI context, that means platform terms, data handling practices, and professional controls must align with confidentiality requirements before client information is entered. The focus should be less on whether AI interactions are discoverable and more on whether confidentiality ever attached. As Heppner shows, confidentiality is not assumed simply because a document resembles legal analysis. It depends on the conditions under which the information was created and maintained.

Reprinted with permission from the March 19, 2026 edition of “The Legal Intelligencer” © 2026 ALM Global, LLC. All rights reserved. Further duplication without permission is prohibited. Request academic re-use from www.copyright.com. All other uses, submit a request to asset-and-logo-licensing@alm.com. For more information visit Asset & Logo Licensing.

Legal Intelligencer: Understanding and Applying Local Rules When Drafting ESI Protocols

Kelly Lavelle — Thu, 20 Nov 2025 21:11:27 +0000

A well-drafted ESI protocol defines production formats, metadata requirements, search terms, and privilege review procedures, reducing disputes and helping discovery move forward efficiently.

In the November 20, 2025 edition of The Legal Intelligencer, Kelly Lavelle writes, “Understanding and Applying Local Rules When Drafting ESI Protocols.“

ESI protocols govern how parties preserve, collect, review, and produce electronic evidence in federal litigation. What once was an obscure procedural topic now lies at the center of nearly every civil case involving documents, emails, databases and other forms of electronic data. ESI protocols provide structure and predictability in an area that can otherwise be costly and difficult to manage. A well-drafted ESI protocol defines production formats, metadata requirements, search terms, and privilege review procedures, reducing disputes and helping discovery move forward efficiently.

The Role of Local Rules in Electronic Discovery

With ESI protocols now a regular part of litigation, local rules and court practices play a more significant role in guiding the process. While the Federal Rules of Civil Procedure provide the foundation for electronic discovery, many district courts have developed their own local rules and forms to guide parties in drafting ESI protocols. Some districts have detailed local rules governing electronic discovery, while others, including the Eastern District of Pennsylvania, have not enacted a formal local rule governing ESI, but rely instead on individual judges’ policies and preferences that serve the same purpose. These guidelines show that courts expect parties to manage ESI in an organized way and to address potential issues in advance rather than waiting for them to arise. Following local practice helps lawyers know what the court expects, makes negotiations smoother, and saves time by avoiding mistakes the court has already ruled out.

Recent case law demonstrates how these local practices guide judicial decisions on ESI. In Hall v. Warren, 2025 WL 1392294 (W.D.N.Y. May 14, 2025), the parties failed to reach an agreement on an ESI protocol governing the production of electronically stored information. When negotiations fell apart, plaintiffs moved to compel production of metadata from internal reports and investigative files, arguing that such information was necessary to verify whether documents had been altered and to confirm their authenticity and completeness. Unable to obtain agreement from the parties, the court issued its own ESI protocol order that included selected provisions from each side’s proposal.

The defendants filed multiple objections to the court’s order, arguing that the order violated the district court’s local rules and imposed undue cost and technical burden. They claimed the order mandated universal metadata production, required technical tasks they could not perform, demanded compatibility with opposing counsel’s document review platform, and ordered native production without proper justification.

The court rejected these arguments, clarifying that the ESI order did not require many of the things defendants claimed it required and that the order’s actual requirements were consistent with the local rules and, in many instances, mirrored the local rules specific provisions. The order also provided flexibility for the parties to meet and confer on alternate formats or cost-sharing if a production became disproportionately burdensome.

The critical lesson from Hall is that the defendants misunderstood both the ESI order and the local rules they said it violated. The defendants clearly had read the district’s local rules because they cited them throughout their objections. Their problem was not ignorance of the local rules but misunderstanding what those rules actually required. The defendants thought the ESI order required things it did not require. They believed it violated local rules when it actually followed them. They argued that certain procedures were mandatory when the order actually provided flexibility. Because they misunderstood the local rules from the beginning, they could not negotiate effectively. As such, the defendants ended up with a court-imposed protocol that likely was less favorable than what they could have achieved through informed negotiation.

While Hall shows how local rules guide substantive resolution of ESI disputes, the case Husidic v. FR8 Solutions, No. 3:24-cv-963-WGY-SJH (M.D. Fla. Sept. 10, 2025), demonstrates the importance of complying with local procedural requirements even when parties agree. In Husidic, the parties submitted a joint motion requesting that the court enter the parties’ stipulated ESI protocol as an order. The magistrate judge denied the motion identifying three defects. First, the motion cited no legal authority for the proposition that courts should routinely enter agreed ESI protocols as orders. Second, the parties failed to include the memorandum of law required by the U.S. District Court for the Middle District of Florida’s local rules governing motion practice. Third, the request lacked any showing that court adoption of the protocol was necessary or appropriate.

The magistrate judge also identified a substantive issue with the parties’ proposed stipulation, which included ambiguous provisions referencing privilege and work-product protections that might implicate Federal Rule of Evidence 502, yet neither the stipulation nor the motion addressed Rule 502(d). The court emphasized that if parties seek a non-waiver order, they must make the request expressly and provide factual support demonstrating why such an order is necessary.

In declining to endorse the parties’ ESI protocol, the court implied that judicial resources should not be spent approving private agreements that can be managed through cooperation and adherence to the local rules. Further, compliance with local procedural requirements is essential, even when parties agree on the substance of an ESI protocol. Counsel must understand not only the local rules governing electronic discovery but also the local rules governing motion practice and the standards for obtaining court orders.

Practical Guidance for Counsel

For lawyers, these cases show that the most effective ESI protocol is one based on the local rules of the district court or specific judges’ policies and procedures. Counsel should familiarize themselves not only with the federal rules but also with the district court’s ESI checklists, policies, procedures, and discovery guidelines, many of which provide practical direction that supplements the federal rules. Before the Rule 26(f) conference, counsel should review the district’s local rules on electronic discovery and consider how those rules, and the judge’s preferences are likely to influence court resolution of any disputes. By following local rules from the beginning, parties can reduce the likelihood of motion practice and avoid judicially imposed resolutions that may be less favorable than a negotiated agreement.

Kelly A. Lavelle is an associate at Kang Haggerty. She focuses on e-discovery and information management, from preservation and collection to review and production of large volumes of electronically stored information. Contact her at klavelle@kanghaggerty.com.

Reprinted with permission from the November 20, 2025 edition of “The Legal Intelligencer” © 2025 ALM Global, LLC. All rights reserved. Further duplication without permission is prohibited. Request academic re-use from www.copyright.com. All other uses, submit a request to asset-and-logo-licensing@alm.com. For more information visit Asset & Logo Licensing.

Legal Intelligencer: Discovery Risks of ChatGPT and Other AI Platforms

Kelly Lavelle — Thu, 21 Aug 2025 16:44:25 +0000

In the August 21, 2025 edition of The Legal Intelligencer, Kelly Lavelle writes, “Discovery Risks of ChatGPT and Other AI Platforms.”

OpenAI CEO Sam Altman recently warned that ChatGPT conversations are not legally protected and can be used as evidence in court. Speaking on a podcast, Altman acknowledged that OpenAI is legally required to retain user chats, including deleted ones, due to a current court order discussed later in this article. Comparing AI conversations to those with doctors, lawyers, or therapists, Altman argued that similar confidentiality protections should exist but currently do not, leaving sensitive exchanges with public AI tools fully exposed to discovery, an issue he described as needing to be addressed with urgency.

The use of AI tools like ChatGPT and Claude has created new issues for the discovery process. Lawyers must recognize that AI queries and outputs may qualify as electronically stored information (ESI) under both federal and state discovery rules. As AI technology becomes more integrated into legal practices, discovery requests are beginning to target the use of these technologies, seeking access to AI-generated documents, search histories and communication logs.

Many users may view AI tools as private assistants rather than potential witnesses. However, the use of AI tools like ChatGPT and Claude can inadvertently expose sensitive information, including legal strategies and privileged facts. Users may not realize that third-party AI platforms can be compelled to produce records during litigation, potentially compromising confidentiality and privilege. Many platforms keep detailed logs that include prompts and generated output, often tied to user accounts. Users may be surprised to learn that if they have employed these tools to draft documents or summarize confidential facts, those entries may be discoverable. This misunderstanding can create serious risks.

Several recent cases illustrate the discovery concerns associated with AI use. In some instances, AI-related ESI has been sought in discovery, challenging both privilege and work product protections. These cases highlight the importance of understanding the discovery implications of AI use and the need for protective measures. The discoverability of ChatGPT searches and their status as non-privileged information in legal proceedings may depend on the context in which the searches are conducted and the applicable privileges.

In some cases, courts have found that AI prompts and outputs, particularly when drafted by counsel, can constitute attorney work product. In Tremblay v. OpenAI, No. 23-cv-03223-AMO, 2024 WL 3748003, at *2-3 (N.D. Cal. Aug. 8, 2024), the court held that prompts created by attorneys reflected their mental impressions and opinions, making them work product rather than mere factual material. Applying that same reasoning, a California district court recently concluded that certain prompts, settings, and outputs from the Claude AI model were likewise protected, rejecting the argument that such materials were automatically discoverable. See Concord Music Group v. Anthropic PBC, 2025 WL 1482734, at *1 (N.D. Cal. May 23, 2025). The court emphasized that these materials were generated during counsel’s investigative process and therefore qualified for work product protection. However, the court also recognized that work product protection can be waived when a party relies on such materials in pleadings or motions. Here, the plaintiffs used specific prompts and outputs in their complaint and preliminary injunction filings, producing nearly five thousand prompt-output records that they relied upon. Under the “fairness principle,” this created a limited waiver. However, the court refused to extend it to all prompts, settings, and outputs, finding that discovery requests for every AI interaction were overreaching and not “closely tailored” to the opposing party’s legitimate needs nor limited to what is necessary under the fairness principle.

On the other hand, and perhaps the most sweeping example is the ongoing New York Times v. OpenAI lawsuit, which alleges that OpenAI unlawfully used millions of Times articles to train its AI models, including ChatGPT. In connection with that case, on May 13, 2025, Magistrate Judge Ona T. Wang ordered OpenAI to “preserve and segregate all output log data that would otherwise be deleted on a going forward basis,” a directive affecting hundreds of millions of ChatGPT users. See In re OpenAI Copyright Infringement Litigation (relating to The New York Times v. Microsoft, 23-cv-11195), No. 25-md-3143 (SHS) (OTW), ECF No. 551, at 2 (S.D.N.Y. May 13, 2025). OpenAI objected, arguing the order forced it to “disregard legal, contractual, regulatory, and ethical commitments” and retain up to 60 billion conversations, of which the plaintiffs estimated only 0.006% might be relevant. See OpenAI Defs.’ Supplemental Opp’n to News Pls.’ Mot. Regarding Output Logs , No. 25-md-3143 (SHS) (OTW) (S.D.N.Y. May 23, 2025), ECF No. 578. The district judge rejected those arguments, affirming the order and noting that OpenAI’s own terms of use allowed preservation for legal requirements. See Order, No. 25-md-3143 (SHS) (OTW), ECF No. 712 (S.D.N.Y. June 26, 2025). Although the court later clarified that certain categories were excluded, the case highlights how deleted AI data can become subject to preservation and discovery.

Lawyers should address these concerns and advise clients at the outset that interactions with AI tools may be discoverable and should be treated accordingly. This means instructing clients not to input privileged or confidential strategy into public or unsecured AI platforms. Clients should understand that the same caution they would apply to an email applies to AI prompts.

Further, the integration of AI tools like ChatGPT into legal practices necessitates a careful consideration of discovery obligations. Firms should adopt clear policies governing the use of AI tools for litigation tasks. These policies should address acceptable uses, data protection measures, and procedures for handling AI-generated content.

Reprinted with permission from the August 21, 2025 edition of “The Legal Intelligencer” © 2025 ALM Global, LLC. All rights reserved. Further duplication without permission is prohibited. Request academic re-use from www.copyright.com. All other uses, submit a request to asset-and-logo-licensing@alm.com. For more information visit Asset & Logo Licensing.

Legal Intelligencer: Revisiting ‘Zubulake’ 20 Years Later

Kelly Lavelle — Thu, 19 Jun 2025 14:21:53 +0000

In the June 19, 2025 edition of The Legal Intelligencer, Kelly Lavelle writes, “Revisiting ‘Zubulake’ 20 Years Later.”

Introduction

It has been 20 years since Judge Shira A. Scheindlin issued the landmark Zubulake decisions, a series of rulings that profoundly reshaped e-discovery practices in federal litigation. At a time when electronically stored information (ESI) was rapidly expanding and began to overwhelm traditional discovery practices, Zubulake addressed critical issues related to preservation, production, and cost-allocation. These decisions expanded the definition of ESI, established new standards for attorney oversight, and set a precedent for holding both parties and their attorneys accountable for failing to fulfill their e-discovery obligations. This article revisits Zubulake and explores its enduring impact on e-discovery standards and practices, as well as the significant developments that have occurred since these pivotal decisions.

The ‘Zubulake’ Decisions—A Brief Overview

The Zubulake decisions refer to a series of five influential rulings between 2003 and 2004 in Zubulake v. UBS Warburg, a gender discrimination case filed by Laura Zubulake in the U.S. District Court for the Southern District of New York. Judge Scheindlin addressed, for the first time, many of the challenges posed by electronic discovery. Collectively, the Zubulake decisions established a framework for managing electronic discovery in federal litigation.

In the first three decisions, Zubulake I-III, the court laid the groundwork for how ESI should be treated in discovery. The rulings addressed issues of accessibility, relevance, and cost-shifting in the context of electronically stored information, providing a framework for assessing the burdens and responsibilities of document production. In Zubulake I, Scheindlin set forth a test to determine whether the costs of producing ESI should be shifted to the producing party or the requesting party. Zubulake II further refined this analysis, focusing on the evaluation of relevance and the burden of inaccessible data. In Zubulake III, the court applied these principles to the specific facts of the case and ordered the defendant to produce the ESI at its own expense. Zubulake I through III provided clarity on how to balance the burdens of e-discovery, especially when dealing with large volumes of data and inaccessible data.

In Zubulake IV, the court shifted its analysis to preservation duties and attorney oversight, detailing when the obligation to preserve arises and what steps lawyers must take to ensure compliance. Scheindlin held that the obligation to preserve ESI arises when litigation is reasonably anticipated, not once a lawsuit is filed. She emphasized that this duty extends not only to the parties but also to their attorneys, who are responsible for ensuring that their clients implement proper litigation holds and suspend routine data destruction practices.

In Zubulake V, the court addressed the consequences of failing to meet e-discovery obligations, making it clear that failures in preservation can result in sanctions, including adverse inference instructions. Scheindlin not only ordered an adverse inference instruction to the jury, instructing the jury that it may presume the missing evidence was adverse to the case of the defendant, but also established the duties of counsel in implementing legal holds and preserving electronic evidence in discovery. The decision put lawyers on notice that the reasonable anticipation of litigation triggering e-discovery preservation obligations often occurs well before litigation is commenced.

Together, these decisions formed the foundation for modern e-discovery practice and directly influenced the 2006 and 2015 amendments to the Federal Rules of Civil Procedure, which codified many of the principles Scheindlin had established.

Impact on E-Discovery Practices

Establishing a Framework for Cost-Shifting in E-Discovery

The most significant impact of the Zubulake decisions was the establishment of a framework for determining when cost-shifting in e-discovery is appropriate. In Zubulake I, the court set forth the seminal seven-factor test to determine whether the costs of producing ESI should be shifted to the producing party or the requesting party. The factors include the extent to which the request is specifically tailored to discover relevant information, the availability of such information from other sources, the total cost of production relative to the amount in controversy, the cost of production compared to the resources available to each party, the relative ability of each party to control and minimize costs, the importance of the issues at stake in the litigation, and the relative benefits to the parties of obtaining the information. This test was designed to simplify the application of the Rule 26 proportionality standard in the context of electronic data and continues to serve as the legal standard in the analysis of cost allocation in e-discovery.

Defining Preservation Duties and Consequences for Spoliation

Another significant contribution of Zubulake, specifically Zubulake IV and V, was the court’s establishment of clear standards governing the duty to preserve ESI and the consequences of failing to do so. These decisions highlighted the importance of implementing effective litigation holds, clarified the point at which the duty to preserve is triggered, and emphasized the obligation of counsel to ensure compliance with preservation duties. The rulings also provided guidance on when sanctions for spoliation are appropriate and affirmed that an adverse inference instruction may be imposed in response to willful destruction of ESI. The opinion shifted the focus of e-discovery from merely producing data to proactively safeguarding it at the initial stages of the dispute.

Developments Since ‘Zubulake’

Technological Advancements

In the years since the Zubulake decisions, the rapid growth of cloud computing, social media, messaging applications and mobile devices has significantly expanded both the scope and complexity of ESI. These technological developments have created new challenges in the identification, preservation and review of e-discovery during litigation. However, despite the ever-evolving nature of e-discovery, the core principles established in Zubulake, such as proportionality, preservation and attorney oversight, remain essential to conducting effective and defensible discovery.

Amendments to the Federal Rules of Civil Procedure

In response to the growing challenges of e-discovery highlighted in Zubulake, the Federal Rules of Civil Procedure were amended in 2006 and again in 2015 to formally incorporate many of the principles established in Zubulake. The 2015 amendments to Rule 26(b)(1) reinforced the concept of proportionality and the obligation of the parties to consider these factors in making discovery requests, responses, or objections. Additionally, Rule 37(e) was revised to address the preservation of ESI and to establish a uniform approach to dealing with spoliation and provides a framework for imposing sanctions, reflecting the influence of Zubulake on modern e-discovery standards.

Emerging Trends

In keeping with the growing emphasis on proportionality and cooperation in e-discovery, courts are increasingly focused on ensuring that discovery efforts are reasonable and proportional to the needs of the case. Additionally, the use of technology-assisted review (TAR) and artificial intelligence (AI) in e-discovery continues to transform the document review process, offering new opportunities for efficiency and accuracy in document review. The tools further advance the principles established in Zubulake by enabling more targeted, cost-effective and defensible discovery practices.

Conclusion

Twenty years later, the Zubulake decisions continue to shape legal practice. They established foundational principles that remain integral to e-discovery standards and procedures. While technological advancements and amendments to the Federal Rules of Civil Procedure have introduced new challenges and opportunities, the core principles of Zubulake remain relevant and a critical component of modern e-discovery. As the legal profession continues to adapt to the evolving landscape of technology and information management, Zubulake will endure as a defining authority in e-discovery.

Reprinted with permission from the June 19, 2025 edition of “The Legal Intelligencer” © 2025 ALM Media Properties, LLC. All rights reserved. Further duplication without permission is prohibited, contact 877-257-3382 or reprints@alm.com.

Legal Intelligencer: Technology-Assisted Review: A Superior Approach in Legal Document Review

Kelly Lavelle — Thu, 26 Oct 2023 19:15:59 +0000

As technology continues to advance, legal professionals should embrace TAR as an invaluable tool in the search for efficient, accurate, and cost-effective legal document review.

In the October 26, 2023 edition of The Legal Intelligencer, Kelly Lavelle wrote, “Technology-Assisted Review: A Superior Approach in Legal Document Review.”

Technology-assisted review (TAR) has changed how lawyers manage and analyze vast volumes of electronic data in the ever-changing landscape of legal document review. Traditionally, search terms have been the preferred method in the e-discovery process. However, TAR is rapidly emerging as a superior alternative, offering numerous benefits over conventional search terms in legal document review.

Technology-assisted review (TAR) is a process of having computer software electronically classify documents based on input from reviewers to expedite the organization and prioritization of the document collection. As reviewers train the software, it learns to identify and highlight relevant information and ensure quality control accurately. TAR allows reviewers to make decisions rapidly by prioritizing the most critical documents. TAR has been accepted by the U.S. courts since the seminal 2012 decision in Da Silva Moore v. Publicis Groupe & M.S.L. Group, 287 F.R.D. 182 (S.D.N.Y. 2012) and is now viewed as black-letter law.

The landscape of technology-assisted review options includes: TAR 1.0, also known as “predictive coding” or “simple active learning,” TAR 2.0, known as “continuous active learning,” TAR 3.0, which is similar to continuous active learning but applied to cluster centers and TAR 4.0 known as hybrid multimodal I.S.T. (intelligently spaced training) predictive coding.

There are several differences between TAR 1.0 and TAR 2.0, which lie in how the algorithm is trained. TAR 1.0 software training begins by taking a random sample of documents from the entire TAR set. A reviewer then codes those documents, and based on the coding in that set (seed set), the software generates a predictive model that is applied across all relevant documents. In contrast, with TAR 2.0—the more advanced approach, the reviewer and software training occur simultaneously, and the algorithms are continuously learning as the documents are reviewed. TAR 3.0 combines the advantages of TAR 1.0 and TAR 2.0. The key aim of TAR 3.0 is to leverage the features of TAR 2.0 with new techniques that will allow for earlier and smarter identification of relevant documents. TAR 4.0 uses a combination of human and machine functions to efficiently and accurately identify relevant data. TAR 3.0 and 4.0 essentially fall under the TAR 2.0 umbrella.

Legal professionals have shifted toward adopting TAR 2.0 due to various advantages. In particular, continuous active learning has been shown to reach higher levels of recall, identifying a greater number of relevant documents more quickly and with less effort by the reviewer than TAR 1.0. Continuous active learning can also readily accommodate changes in the scope of discovery and rolling document productions since it continues training throughout the review process. A TAR 1.0 algorithm stops training when it reaches stable quality results, regardless of how many documents are subsequently reviewed, and requires manual re-training after each document review cycle is complete. A TAR 2.0 algorithm is trained by every coding decision until the review stops. Because the TAR 1.0 algorithm is fully trained before the review begins, it does not adapt to changes in the scope of discovery, such as the addition of documents in a rolling production or the addition of legal issues involved.

Determining which protocol best fits a particular matter depends on case objectives and requires a more detailed understanding of the various methodologies. Whichever variation is used, there are a number of key benefits to using TAR in document review.

Enhanced Precision and Recall

One of the primary advantages of TAR is its ability to enhance both precision and recall significantly. In the context of TAR, precision is a measure of how often an algorithm accurately predicts a document to be responsive. Essentially, it measures the percentage of documents produced that are actually responsive. Recall is a measure of completeness, referring to the percentage of relevant documents identified within the entire universe of documents.

Traditional search terms rely on specific keywords or phrases, potentially missing relevant documents that do not contain those exact terms. TAR, however, uses sophisticated machine-learning algorithms to identify document patterns. This advanced approach enables TAR to uncover relevant documents that might have been overlooked relying solely on search terms.

Reduced Volume of Irrelevant Documents

Search terms often generate large sets of irrelevant documents, burdening reviewers with time-consuming manual review. TAR minimizes this burden by prioritizing the review of documents most likely to be relevant. TAR learns from the coding decisions applied by the reviewers, prioritizes the most likely relevant documents, and excludes irrelevant ones. For projects where documents do not need to be reviewed before production, technology-assisted review can quickly separate the documents into relevant and irrelevant categories without review, allowing for quick production and substantial time and cost savings.

Improved Consistency and Defensibility

TAR offers a consistent and defensible approach to document review. From the initial approval of its use in Da Silva Moore, TAR has become a staple in modern litigation. Unlike search terms, which can be ambiguous and susceptible to implementation errors, TAR’s machine-learning algorithms apply consistent criteria across all documents. This uniformity not only enhances accuracy of the review process but also bolsters its defensibility in legal proceedings, as the process is well-documented and transparent.

Time and Cost Efficiency

Numerous courts have recognized the cost-saving benefits of technology-assisted review. By using TAR, legal professionals can significantly reduce review time, eliminating the hours spent on manual document review and reducing attorney fees substantially. Further, TAR provides the advantage of early case assessment by prioritizing relevant documents, thereby reducing the overall volume of documents. This proactive approach provides attorneys with the necessary information to make informed decisions about case strategy, settlement negotiations, or the need for additional evidence. Ultimately, TAR saves both time and valuable resources.

The advantages of technology-assisted review over traditional search terms in the modern legal landscape are clear and compelling. TAR not only offers enhanced precision but also drastically reduces the amount of irrelevant documents in the review set. TAR provides improved consistency coupled with significant cost savings. Its adaptability and early case assessment capabilities further solidify its role as a superior approach to document review in e-discovery. As technology continues to advance, legal professionals should embrace TAR as an invaluable tool in the search for efficient, accurate, and cost-effective legal document review.

Reprinted with permission from the October 26, 2023 edition of “The Legal Intelligencer” © 2023 ALM Media Properties, LLC. All rights reserved. Further duplication without permission is prohibited, contact 877-257-3382 or reprints@alm.com.