Back to Blog

The Privilege Puzzle in eDiscovery: Untangling the Mystery

August 1, 2023
The Privilege Puzzle in eDiscovery

The legal right to withhold privilege information is a blessing for organizations and a headache for eDiscovery practitioners.  In order to protect an organizations privilege content, practitioners must first identify it.  For over a decade, identifying privileged content has been manual (human doc review) and excessively time-consuming.  Many practitioners have turned to advancements in machine learning to help automate the privilege identification process.  Some have moved beyond the review phase and are bringing efficiency and cost reduction to their clients by automating their privilege logs.

In this article, we will discuss the concept of privilege, why managing privilege content is so important, and how teams are leveraging modern techniques to improve efficiency.

Types of Privilege

The FRCP section 26(b)(5) provides guidance on the type of content that may qualify a record to be withheld.

  • Attorney-client privilege: Communication between the attorney and their client.
  • Attorney Work-product privilege: It contains documents prepared in anticipation of litigation, document set for trial preparation, reports, correspondence, and other documents
  • Privacy Laws Privilege: Documents that can be considered as “confidential” under the right to privacy by data privacy laws or any other statutory laws.
  • Trade Secrets Privilege: Documents that contains formulas, product designs, patent, business models, or processes. All Intellectual Property (IP) information that can be leveraged by the opposing party if revealed.

Why is it Important to manage privilege?

Google LLC v. Oracle America, Inc., No. 18-956, (US, Apr. 5, 2021) is an example to understand the importance of managing privilege documents during eDiscovery.

In the Oracle America, Inc. v. Google, Inc. case, the court addressed the issue of privileged documents inadvertently produced by Google. The dispute revolved around an email authored by Tim Lindholm, a Google software engineer, which became the subject of contention. Google had collected multiple drafts and two final versions of the email, collectively known as the Lindholm Email. The court reviewed these documents and found that Google failed to demonstrate that attorney-client privilege or the work product doctrine protected them.

The Lindholm Email contained discussions about technical alternatives to Java for Android and Chrome and mentioned the need to negotiate a license for Java. The final version of the email was marked as “Attorney Work Product” and “Google Confidential.” Oracle contested Google’s claims of attorney-client privilege and work product protection for the Lindholm Email and sought the court’s intervention to compel the production of the documents.

The court explained that attorney-client privilege protects confidential communications between attorneys and clients made for the purpose of legal advice. However, the privilege is narrowly and strictly construed, and the party asserting it bears the burden of proving its application. In this case, Google failed to meet the requirements of the eight-part test for attorney-client privilege. The court concluded that the privilege did not extend to the Lindholm Email and ordered Google to produce the documents.

This incident highlighted the importance of properly assessing and protecting privileged documents during the discovery process and emphasized the need for parties to exercise caution to avoid inadvertent disclosures that could impact the course of litigation.

How are teams improving their privilege management processing using their eDiscovery Platforms?

  • Building Reusable Predictive Coding Models for Privilege

Identifying privileged content can be a moving target for many teams.  Sometimes requiring background on each particular case and specialized training for each review team, per case.  For teams representing the same company with the same privilege concerns in many cases, AI and predictive coding models offer the opportunity to invest in creating efficiency for the next case.

The efficiency is created by the portability of these models.  Teams can build models to identify one category of privilege, reuse the model across cases, but most importantly hone its accuracy for organization-specific content and privilege concerns.

  • Leveraging Aliasing and Name Normalization

Name normalization groups similar names and variations of names together to eliminate duplicates.  You may find one individual has many email addresses, but in the privilege log they should appear with one identifier.  Their name.  Name normalization allows support teams to “normalize” all of these different emails with one alias so,, and all become John Doe.

This type of normalization and aliasing can be done for any type of personally identifiable information or metadata like a phone number, email, or even a username (for standardizing and anonymizing the attorney’s name who made a particular coding decision).

  • Introducing Templatization and Automation to their Privilege Logs

Privilege logs typically consist of an Excel or spreadsheet noting contextual information (metadata) about a privilege document and a specific reason for withholding it.  In large or complex cases these reports will also include fields to show which reviewer made the decision to withhold and why.  These reports can be cumbersome to manage and are often built manually.  Human error and inadvertent disclosure of the information is not a primary risk in this situation, the real killer is time and cost.  Building these logs typically takes a senior resource multiple hours every day to keep up with basic tasks like field merging and name normalization.  Modern teams are looking to software to automate these repetitive tasks and introduce new ways to have machines build human-readable privilege descriptions.

The mitigation steps didn’t work, what next?

FRE 502 (d): The last resort for protecting privilege

Fortunately, there exist FRE 502(d) law to rescue privilege document in times of glitches. Rule 502(d) order helps you claw back any privilege document unintentionally produced. As per federal law, to avoid the waiver of privilege and raise questions over the costly review process, one must have a 502(d) order. It permits the federal court to issue an order stating that the produced document is protected with client-attorney privilege or trade secret privilege and cannot be used in court.

Similarly, there is 502(b) that comes into picture in case the attorney does not have a 502 order in hand. Here the only change that applies is that one must produce relevant facts to prove if the produced documents they wish to revoke are actually privilege documents. It raises questions on the entire eDiscovery process and puts you in a position of low confidence and credibility.

Most experienced legal professionals advocate on using this order beforehand as it is useful in times of need.


In conclusion, the concept of privileged documents and the management of privileged content holds significant importance in the legal and corporate world. Safeguarding privileged information ensures the confidentiality and integrity of sensitive communications between attorneys and clients, as well as attorney work products. The inadvertent production of privileged documents can have far-reaching consequences, potentially impacting the fairness of legal proceedings and the rights of parties involved, as exemplified in the above mention Oracle v Google case.

With growing legal tech and recognizing the significance of managing privilege, teams are increasingly leveraging modern techniques to enhance efficiency. Advanced technologies, such as AI-powered document review systems and sophisticated data analytics, play a crucial role in identifying and protecting privileged content. These solutions assist in the accurate and efficient identification of privileged materials, reducing the risk of inadvertent disclosures and ensuring compliance with legal and ethical obligations.

By leveraging modern techniques, legal teams can effectively navigate the complexities of privilege, reduce the risk of inadvertent disclosures, and uphold the integrity of legal proceedings.

As the legal landscape continues to evolve, the effective management of privileged documents remains an ongoing challenge. However, by embracing modern techniques and adopting proactive strategies, organizations can navigate this complex terrain with greater efficiency, ensuring the protection of privileged content while maintaining the highest standards of professionalism and legal compliance. We at Knovos have been instrumental in helping our clients to manage their privilege documents efficiently and reduce the risk of inadvertent submission of privilege during discovery.


Connor Morgan
Connor Morgan is a Director of Customer Success Group for Knovos. He brings extensive experience delivering tailored legal technology solutions for eDiscovery, Legal Case Management, and Information Governance. Connor has specialized knowledge working with AmLaw100 Firms, U.S. Federal Government Agencies, and State/Local Governments, to analyze real-world business problems and implement Knovos technology solutions.