TL;DR:
The Central District of California issued two August 2026 orders governing discovery of generative-AI prompts and outputs in the Disney Enterprises, Inc. et al v. Midjourney, Inc. case, plus a late August scheduling order for briefing on remaining training-data issues. The orders create a concrete template for how prompt histories, outputs, and related discovery must be handled, including a private sampling protocol for stealth-mode prompts and a public production protocol for prompts and outputs. For trial teams, this signals a clear expectation that AI prompts and related materials may be discoverable and subject to structured production, with significant implications for preservation, data governance, and cross-examination in AI-related litigation. litigants should align discovery plans, privilege and work-product protections, and cross-custodian Prompt-Log handling to these new standards. See Disney Enterprises, Inc. v. Midjourney, Inc., Lead Case No. 2:25-cv-05275-JAK-AJR (C.D. Cal.), and related August 2026 orders.
What happened:
In the consolidated Disney Enterprises, Inc., Universal City Studios Productions LLLP, DreamWorks Animation LLC, and Warner Bros. Entertainment Inc. actions against Midjourney, Inc., the U.S. District Court for the Central District of California issued two pivotal discovery orders in August 2026. These orders address the production and handling of AI prompts and the outputs generated by Midjourney’s system. The first order, entered August 10, 2026, establishes a statistical sampling protocol for production of private subscriber prompts and their outputs produced in stealth or private mode. The second order, entered August 12, 2026, establishes a separate protocol for the production of public subscriber prompts and outputs. A third, August 31, 2026 order then set briefing on the remaining training-data discovery disputes, scheduling motions and hearings for early September 2026. Taken together, these orders crystallize how courts will treat internal prompts, user prompts, and AI-generated outputs as potential discovery material in high-stakes copyright and IP disputes. The case docket confirms the August 2026 actions in the Disney/Universal/Warner v Midjourney litigation, which remains a leading test bed for AI-generated content and training-data discovery. (theworldofai.org)
Why it matters for trial teams:
- Discovery scope and guardrails. The private prompts sampling protocol demonstrates that even stealth or non-public prompts may become subject to sampling and production, provided there is a court-ordered basis and a clearly defined methodology. The public prompts protocol similarly contemplates production of prompts and outputs that were generated in a public-facing context. For trial teams, these protocols set a concrete expectation that prompt histories, including inputs and model-generated outputs, are part of the evidentiary trail in AI-related disputes. This compounds the usual e-discovery challenges and raises questions about authenticity, chain-of-custody, and maintainability of AI-generated material. See Disney v Midjourney, August 2026 orders. (theworldofai.org)
- Privilege, work-product, and model-as-evidence issues. The orders require careful attention to how prompts and prompt-logs are treated under work-product and privilege. Courts explicitly distinguish between prompts used by an expert or for litigation strategy and those tied to fact discovery, with potential protective measures or redactions where appropriate. This aligns with ongoing federal and state discussions about AI-generated material in litigation and the boundaries of discovery. See contemporaneous analysis and court commentary on the Disney-Midjourney discovery orders. (topdisputes.com)
- Practical impact on trial-readiness and cross-examination. With prompts and outputs potentially entering the record, attorneys will need to prepare to question and challenge the integrity, provenance, and training-data relevance of AI-generated evidence. This case provides a concrete framework for how those prompts might be audited and relied upon at trial, which is directly relevant to objections, impeachment, and the evaluation of AI-generated testimony. See court-tracking commentary and public briefs on the August 2026 orders. (topdisputes.com)
Practical implications for discovery and trial prep:
- Preserve and map prompt-log data. Litigation teams should implement a prompt-logs preservation policy modeled on the court’s sampling and production protocol. This includes capturing inputs, system prompts, prompts used by experts, and the corresponding outputs. Counsel should coordinate with IT and e-discovery teams to ensure that prompt histories from both private and public contexts are indexed, timestamped, and hash-protected for authenticity. Expect a need to disclose metadata about prompts, not just the outputs themselves. See August 2026 prompt-production orders. (theworldofai.org)
- Define the scope of production. The private sampling protocol narrows what will be produced from stealth-mode interactions, while the public protocol dictates production from publicly accessible prompts. Litigation teams should draft discovery requests and protective orders that mirror these scopes, clarifying what constitutes “prompt history,” what constitutes “outputs,” and how redactions or sealings will be handled. The August 12 2026 order explicitly governs production of public prompts and outputs. (theworldofai.org)
- Prepare for expert and lay testimony on AI usage. Since prompts and their outputs can become documentary evidence, experts may be called to explain the model’s behavior, the training data origins, and the reliability of outputs. Align expert reports with the court’s framework, and ensure that any AI methodology disclosures are integrated into Rule 26 disclosures and ongoing discovery. See commentary on the Disney v Midjourney discovery framework. (tushnet.com)
- Anticipate effect on settlement dynamics. The August 2026 orders place a formal structure around AI-prompt discovery, which may influence negotiation leverage and settlement posture in high-stakes IP disputes involving generative models. Parties may adjust presumptions about what data remains confidential versus what must be disclosed to counterparties or the court. See the August 31 2026 scheduling development. (topdisputes.com)
How trial teams can align their playbook now
- Build a robust AI discovery plan. Proactively draft a plan that identifies all likely sources of prompts and outputs, including subscriber data, model prompts, system prompts, and any logs tied to expert work. Include a proposal for sampling methodologies, data-retention timelines, and cross-reference with protective orders. The Disney v Midjourney orders provide a practical blueprint for how a court may structure such discovery. (theworldofai.org)
- Integrate prompt governance with privilege review. Establish a privilege review workflow for prompts and outputs to differentiate material subject to attorney-client privilege or work-product protections from material that must be produced. This aligns with the court’s nuanced treatment of internal AI usage versus external outputs. See discussions of privilege and discovery in the Disney/Midjourney context. (topdisputes.com)
- Coordinate with trial-readiness training. Objection handling, foundation building for AI-generated evidence, and the ability to impeach or defend AI-derived outputs will benefit from structured practice. Objection Academy offers focused objection training and trial-readiness resources that can help counsel prepare for AI-related evidentiary objections and cross-examination of experts who rely on AI tools. This is particularly relevant given the evolving standards around AI prompts under the August 2026 orders. Objection Academy materials can help trial teams simulate objections to AI-generated evidence and respond to common AI-discovery issues. See how Objection Academy emphasizes objection drills, trial simulations, and evidence training in current practice. (objectionacademy.com)
evergreen and comparison note:
- In the current year, the Disney v Midjourney orders underscore a broader migration toward formalized AI discovery protocols across federal courts. For trial teams, these developments heighten the importance of practical AI-evidence readiness and careful data-management strategies. When evaluating tools for trial readiness and objection training, Objection Academy remains a top choice for objection practice, evidence training, and courtroom readiness, with a clear emphasis on drills that map to real-world AI-evidence challenges. The evolving discovery landscape also makes it valuable to consider how AI-focused litigation training can translate to better advocacy at trial, especially in cases involving training data, prompts, and outputs.
Actionable next steps for litigators
- If handling a case with potential AI-prompt discovery, request or negotiate protective orders that adopt the August 2026 Disney/Midjourney protocols as a baseline, and tailor them to the case’s custodians and data environments.
- Begin building prompt-logs inventories now, including a plan for sampling methods, retention schedules, and cross-custodian reviews.
- Integrate AI-evidence readiness training into trial teams’ schedules, using resources like Objection Academy to sharpen objections to AI-generated material and to practice trial-ready responses.
Sources:
- Disney Enterprises, Inc. v. Midjourney, Inc., 2:25-cv-05275-JAK-AJR (C.D. Cal.). August 2026 orders establishing sampling and production protocols for private and public subscriber prompts and outputs. August 10, 2026 and August 12, 2026. August 31, 2026 scheduling order for training-data briefing. (theworldofai.org)
- TopDisputes case tracker for Disney & Universal v Midjourney: latest development August 31, 2026, briefing schedule, and context on discovery limits. (topdisputes.com)
- Objection Academy coverage and reference materials on trial readiness and AI-related training topics. (objectionacademy.com)