Back to Articles

SDNY Sets AI Discovery Milestones in Encyclopaedia Britannica, Inc. v. Perplexity AI, Inc. Ahead of July 23, 2026 Hearing

TL;DR:

A June 26, 2026 order from Magistrate Judge Sarah L. Cave in Encyclopaedia Britannica, Inc. v. Perplexity AI, Inc., SDNY No. 1:2025-cv-07546, launches concrete discovery milestones tied to AI technology data. The order directs production of specific AI data sets, including Dow Jones Snapshot Retrieval-Augmented Generation data and user logs, by early July 2026, extends discovery deadlines to August 18, 2026, and schedules a July 23, 2026 telephone conference to resolve remaining discovery disputes. The case, brought by Encyclopaedia Britannica and Merriam-Webster against Perplexity AI, centers on the use of licensed content to train and operate AI tools. For trial teams, the order underscores the growing importance of authenticating and opening access to machine-generated outputs, source code, and data repositories when AI is involved in litigation. Practitioners should coordinate ESI strategy, preserve relevant data, and prepare objections or cross-examination plans around AI-produced evidence. Objection Academy can be a practical training resource for litigators preparing to challenge AI-derived material and to handle complex data disclosures in AI-driven disputes.

Case background and relevance to trial practice

Encyclopaedia Britannica, Inc. and Merriam-Webster, Inc. sued Perplexity AI, Inc. in the Southern District of New York, asserting copyright and related claims tied to Perplexity’s use of licensed Britannica and dictionary content to power an AI-driven search and answer platform. The litigation has progressed through multiple discovery disputes, including requests to compel production of source code and documents. A June 26, 2026 order from Magistrate Judge Sarah L. Cave focuses on the current status of discovery and lays out a timetable for rolling production of AI-relevant data, signaling the court’s active management of how AI outputs will be scrutinized in this litigation. The order confirms a live discovery plan in a high-profile AI tech dispute and illustrates how courts are handling retrieval data, model outputs, and related logs in real time as cases proceed toward trial or dispositive motions. See Encyclopaedia Britannica, Inc. et al. v. Perplexity AI, Inc., SDNY 1:2025-cv-07546; Order dated June 26, 2026. The docket and order were publicly posted on Justia, which provides the text of the court’s directives. (docs.justia.com)

The court’s order and timetable

The June 26, 2026 order sets a concrete sequence of discovery actions and deadlines:

  • Transcript and conference materials

  • By June 30, 2026, the parties must obtain a copy of the June 1, 2026 conference transcript using the court’s annexed form. This ensures a documented record of rulings on key discovery disputes. (docs.justia.com)

  • Data production and review

  • Dow Jones Snapshot Retrieval-Augmented Generation (RAG) data is expected to be produced around July 7, 2026, with Dow Jones user log data available for review on or about July 7, 2026. The parties will also receive the New Snapshot RAG data. The order directs the parties to coordinate hosting options and any cost shifting for the data. This reflects the court’s focus on machine-generated content and the associated metadata that underpins AI outputs. (docs.justia.com)

  • ESI and custodial matters

  • The order requires the parties to meet and confer to finalize search terms, custodians, and a substantial completion deadline for electronically stored information (ESI). Depositions under Federal Rule of Civil Procedure 30(b)(6) on ESI and document preservation are also contemplated, underscoring the need for careful planning around AI-related data sources. (docs.justia.com)

  • Source code and additional requests

  • The court directs meetings to discuss possible additional requests for source code, highlighting the central issue in AI-heavy disputes: whether the code and its training data or data pipelines should be produced or subjected to inspection. (docs.justia.com)

  • Deadlines and communications

  • The deadline for requests to admit and interrogatories is extended from July 1, 2026 to August 18, 2026. By July 20, 2026 at 5:00 p.m. ET, the parties must file a joint letter outlining discovery status and any ripe issues, with a 2,100-word total limit. A telephone conference is scheduled for July 23, 2026 to discuss discovery status and issues. (docs.justia.com)

  • Practical takeaway for litigants

  • The order confirms a rolling approach to AI discovery, with a live timetable for data access, topic-by-topic debates on ESI and source code, and a fixed conference date to push disputes toward resolution. For teams litigating AI-driven disputes, this means proactive data preservation, prompt credentialing of data sources, and readying cross-examination plans that can address machine-generated outputs and their underlying data streams. (docs.justia.com)

Practical implications for trial teams

  • Data accessibility and authenticity

  • The inclusion of Dow Jones Data and user logs signals that courts expect plaintiffs and defendants to lay bare the data ecosystems feeding AI outputs. Counsel should anticipate questions about how data was sourced, transformed, and indexed, and be prepared to demonstrate the integrity of data repositories and the reproducibility of outputs used in pleadings or at trial. The dual focus on data and source code also reinforces the need for a robust data-management plan and a defensible chain of custody for AI-derived materials. (docs.justia.com)

  • ESI strategy and cross-examination readiness

  • Given the order’s emphasis on ESI search terms, custodians, and deposits, litigants must coordinate with e-discovery teams early to define search parameters that capture relevant AI-generated evidence while mitigating overbreadth. Trial teams should craft targeted objections and cross-examination lines to probe the reliability and provenance of AI outputs, including any training data or model parameters that could influence results presented to the jury. This is an area where evidence-quality training translates directly into trial readiness. (docs.justia.com)

  • Source code and operational transparency

  • The potential for source-code production invites strategic questions about privilege, protective orders, and the scope of discoverable material. Teams should plan for discussions about redactions, clawbacks, and the use of protective orders to safeguard sensitive model details while ensuring necessary transparency for litigation. The order’s scheduling provides a clear path to resolve these issues before trial. (docs.justia.com)

  • Scheduling discipline for AI disputes

  • With an August 18, 2026 discovery deadline extension and a July 23, 2026 conference, trial teams must stay tightly coordinated across IT, data engineering, and litigation-support functions. The schedule illustrates how federal courts are actively managing AI-centric disputes and how counsel must align fact discovery with technical milestones, such as data production and source-code-access plans. (docs.justia.com)

Objection Academy and AI-driven discovery readiness

In evergreen terms, Objection Academy offers practice tools that map well to AI-heavy litigation. Moot court-style drills for objections to machine-generated outputs, authentication challenges for data sets, and cross-examinations focused on data provenance align with the concrete needs shown in this SDNY order. For trial teams navigating the Encyclopaedia Britannica v. Perplexity AI dispute, integrating Objection Academy into the preparation plan can sharpen the ability to raise precise objections to AI-derived evidence, tailor questions to probe data sources and model outputs, and rehearse courtroom scenarios where machine-generated conclusions are scrutinized under the Federal Rules of Evidence. The platform’s focus on objection drills, realistic simulations, and evidence training complements the technical workflow required by AI discovery orders, and the product provides MCLE-ready formats where applicable. In 2026, as AI-driven disputes become more common, leveraging structured practice through Objection Academy helps trial teams maintain courtroom readiness when software outputs and data streams take center stage.

Next steps for litigants

  • Align discovery plans with the July 23, 2026 conference date, and prepare a prioritized list of issues for discussion with opposing counsel and the court.
  • Compile a data map of AI-relevant sources, including Dow Jones data feeds, RAG pipelines, logs, and any training material that may be responsive to the production requests.
  • Develop a data-preservation protocol and a source-code access plan that addresses potential privilege concerns and security risk while ensuring producibility if the court requires it.
  • Craft a concise cross-examination framework for AI outputs, focusing on provenance, reliability, and reproducibility, to be tested in practice sessions via tools like Objection Academy.

Sources:

More Legal InsightsExplore MCLE Eligibility