Andersen v. Stability AI Ltd.
Andersen v. Stability AI Ltd.
2025 WL 870358 (N.D. Cal. 2025)
March 19, 2025
Cisneros, Lisa J., United States Magistrate Judge
Summary
The court has adopted a modified version of the parties' proposed ESI protocol, emphasizing the importance of cooperation and specific guidelines for ESI discovery. The court has addressed issues such as search terms, hit reports, and validation methods, and has approved the addition of protocols for enterprise messaging and mobile device data. The court has also directed the parties to use reasonable steps for deduplication and has granted Plaintiffs' requests for handling embedded files and hyperlinked documents.
SARAH ANDERSEN, et al., Plaintiffs,
v.
STABILITY AI LTD., et al., Defendants
v.
STABILITY AI LTD., et al., Defendants
Case No. 23-cv-00201-WHO (LJC)
United States District Court, N.D. California
Filed March 19, 2025
Cisneros, Lisa J., United States Magistrate Judge
ORDER ON ESI PROTOCOL; PROTECTIVE ORDER; DEPOSITION PROTOCOL
I. INTRODUCTION
*1 Plaintiffs, a group of visual artists whose works were allegedly used to train a generative artificial intelligence (AI) model called Stable Diffusion, brought this putative class action on behalf other similarly situated artists against Defendants, five companies that created and/or used Stable Diffusion. Among other claims, Plaintiffs allege that Defendants have engaged in widespread copyright infringement. See generally ECF No. 238 (Second Amended Complaint). This matter was referred to the undersigned for discovery. ECF No. 250. The parties have been unable to agree on the terms of a stipulated electronically stored information protocol (ESI Protocol) and stipulated protective order (Protective Order) and have submitted competing proposals for both. The parties have also been unable to agree on the number of depositions of Defendants' fact witnesses that Plaintiffs may conduct.
II. ESI PROTOCOL
Judge Orrick's Standing Order specifies that parties who seek an order on ESI discovery protocols “shall use” one of the Northern District's models, “unless good cause exists to depart from the model order.” Standing Order ¶ 5. The parties have agreed to use the Northern District's Model Stipulated Order Regarding Discovery of ESI (Model ESI Order) as a starting point and adapt it to fit their needs but disagree about the appropriate extent of modifications. In some instances, Defendants have advocated for certain language from the district's Model Stipulated Order re Discovery of Electronically Stored Information for Patent Litigation (Patent Model ESI Order). Given the nature of the lawsuit, the Court anticipates that there will be a large volume of ESI in diverse forms. A detailed ESI protocol, similar to what Plaintiffs have proposed, is warranted. See Harden v. Mendocino Coast District Hospital, 17-cv-05554, 2019 WL 4256383, at *1 (N.D. Cal. Sept. 9, 2019) (“[A]n ESI protocol has to be specific or it doesn't mean anything.”).
A. Paragraphs 1 and 2
The Court rejects Defendants' proposed addition to Paragraphs 1 and accepts Defendants' proposed language to Paragraph 2 that incorporates language from this district's Model Patent Order. Thus, Paragraphs 1 and 2 shall read:
1. PURPOSE
This Order will govern discovery of electronically stored information (“ESI”) in this case as a supplement to the Federal Rules of Civil Procedure, this Court's Guidelines for the Discovery of Electronically Stored Information, and any other applicable orders and rules.2. COOPERATIONThe parties are aware of the importance the Court places on cooperation and commit to cooperate in good faith throughout the matter consistent with this Court's Guidelines for the Discovery of ESI. As in all cases, costs may be shifted for disproportionate ESI production requests pursuant to Federal Rule of Civil Procedure 26. Likewise, a party's nonresponsive or dilatory discovery tactics are cost-shifting considerations.
B. Paragraph 4
The Court accepts the parties' largely agreed upon Paragraph 4. See ECF No. 248-3 ¶ 4. As to Defendants' modification to Plaintiffs' addition to Paragraph 4, the Court expects that the “reasonable steps” the parties will take to implement litigation holds include establishing procedures to make sure that notices of the holds are easily understood, and received by the necessary parties and acted upon. See ECF No. 249-4 ¶ 4. Accordingly, Paragraph 4 has been modified slightly for clarity and shall read:
*2 The parties have an obligation to take reasonable steps to preserve discoverable information in the parties' possession, custody, or control, which includes Metadata where applicable, consistent with the Federal Rules.
The parties that are business entities have issued, or immediately after this Order has taken effect shall issue, litigation hold notices to those internal departments, divisions, committees, teams, and individuals likely to possess potentially responsive information, and to persons or entities responsible for maintenance of non-custodial sources, likely to contain potentially responsive information. These business entities shall establish procedures, if they have not already done so, to ensure that the notices are easily understood, and have been received and appropriately acted upon.
C. Paragraph 5 and its subparagraphs
Plaintiffs and Defendants have presented competing proposals under the headings “Search” and “ESI Search,” respectively. Plaintiffs propose a series of five subparagraphs that address disclosures, search methodology, search terms, hit reports, and validation. Defendants propose a single paragraph that focuses on the selection of search terms and search strings. The Court adopts in part Plaintiffs' approach because it sets parameters for a series of topics related to ESI search that will clarify and advance cooperation for purposes of ESI discovery.
Plaintiffs' first proposed subparagraph, Paragraph 5.a is adopted to the extent it is consistent with this district's Guidelines for the Discovery of Electronically Stored Information (ESI Guidelines). The ESI Guidelines encourage the parties to discuss certain ESI-related topics as part of the required Rule 26(f) meet and confer conference, including the sources, scope, and type of ESI, as well as date ranges and other details. See ESI Guideline 2.02.
Plaintiffs' proposed Paragraph 5.a also requires the disclosure of certain preservation-related information in writing, namely the specific reasons why categories or sources of discoverable information have not been preserved or should not be preserved. In contrast, Defendants' proposed protocol does not address disclosures or discussions regarding the scope of preservation. It is well recognized that early discussion of the scope of preservation can reduce disputes and the need for court intervention. See Sedona Principles, Third Edition, Principle 3 (“As soon as practicable, parties should confer and seek to reach agreement regarding the preservation and production of electronically stored information.”); Comment 3.a. (noting that parties may seek to resolve early the identification of data sources that will be subject to preservation, and the relevant time period, among other issues). The Court modifies Paragraph 5.a to align it with the ESI Guidelines' recommendations and Rule 26(b)(2)(B) of the Federal Rules of Civil Procedure. At the outset of the discovery process, this protocol does not require Defendants to detail in writing the specific reasons why categories or sources of discoverable information have not been preserved or should not be preserved. See ESI Guideline 2.01(e).
The parties also disagree on the number of custodians. Plaintiffs propose that each party identify the twenty-four custodians believed to most likely have discoverable ESI, while Defendants' counterproposal provides that “the parties will identify up to five (5) custodians believed to most likely have discoverable ESI.” In this action there are eight individual named Plaintiffs[1] and multiple Defendants: Stability AI Ltd. and Stability AI, Inc. (collectively “Stability AI”), Midjourney, Inc., DeviantArt, Inc., and Runway AI, Inc. On one hand, given the number of individual named Plaintiffs, Plaintiffs' proposal raises the prospect of a significant, perhaps excessive number of custodians relative to the number and size of each Defendant and the issues involved in this case. At the same time, it is not clear from the record the number of employees at each defendant company, and whether each Defendant is a similarly sized company. On the other hand, Defendants' counter proposal appears overly restrictive. To resolve the number of custodians, the Court directs that a Requesting Party may request up to twelve custodians per Producing Party. Named Plaintiffs are collectively considered a single “Requesting Party,” and “Producing Party.” See Nazemian v. NVIDIA, 24-cv-01454-JST, Order Regarding Discovery of Electronically Stored Information (Nazemian ESI Order), ECF No. 98 at 3. Stability AI Ltd. and Stability AI, Inc., collectively Stability AI, is considered a single Requesting and Producing Party.
*3 Thus, incorporating the Court's modifications discussed above, Paragraph 5.a, regarding disclosures shall state:
Consistent with the Parties' obligations under the Rules, the Parties will meet and confer to plan for discovery. This will include a disclosure of custodians, data sources, anticipated date ranges of relevant documents, and categories of documents or ESI that contain information relevant to this litigation.
The Parties shall endeavor to reach an agreement regarding the data sources subject to preservation and discovery. To advance that effort, the Parties shall disclose (a) categories or sources of relevant information that are not reasonably accessible because of undue burden or cost, and (b) categories of data or sources that a party believes could contain relevant information, but are not discoverable and should not be preserved under the proportionality factors. A Requesting Party may request up to twelve custodians per Producing Party. Named Plaintiffs are collectively considered a single “Requesting Party,” and a single “Producing Party.” Stability AI Ltd. and Stability AI, Inc., collectively Stability AI, is considered a single Requesting and a single Producing Party. The parties may jointly agree to modify these limits without the Court's leave. The Parties may request inclusion of additional custodians as discovery progresses. The Court shall consider contested requests for additional custodians, based upon a showing of proportionality to the needs of the case as that Phrase is used in Rule 26 of the Federal Rules of Civil Procedure.
Next, Plaintiffs' proposed Paragraph 5.b. addresses search methodology by requiring a Producing Party to disclose certain information, namely search terms, keywords, date limitations and custodians, prior to using a search methodology to identify documents for production. ECF No. 249-4 at 4-5. Defendants' proposed protocol does not contain an equivalent requirement. See ECF No. 248-3 at 4. The parties, however, have agreed that in responding to a request for production of documents under Federal Rule of Civil Procedure 34, they will “meet and confer about methods to search ESI[.]” Id.; ECF No. 249-4 at 4.
Plaintiffs' proposed Paragraph 5.b does not dictate what methodologies a Producing Party must use or authorize the Requesting Party to choose the search methodolog(ies). The paragraph contemplates that by being informed of certain aspects of the search method(s), the Requesting Party will be able to independently evaluate the search method being used to collect ESI before it is deployed. The mere disclosure of this information does not run afoul of the Sedona principle observed in NuVasive, Inc. v. Alphatec Holdings, Inc., No. 18-cv-0347, 2019 WL 4934477, at *2 (S.D. Cal. Oct. 7, 2019), which recognized that the responding party is generally best situated to evaluate procedures and methodologies appropriate for producing their own ESI. Id. The Court's concern, however, is that the strict timing for the disclosure—prior to a party's use of a search method—is an unnecessary, preemptive restraint on responding parties.
*4 The Sedona Conference is a source for guidance related to ESI discovery, and its commentary related to this issue is nuanced. On one hand, the Sedona Conference's Comment 6.b explains, “as a general matter, neither a requesting party nor the court should prescribe or detail the steps that a responding party must take to meet its discovery obligations, and there should be no discovery on discovery, absent an agreement between the parties, or specific, tangible, evidence-based indicia (versus general allegations of deficiencies or mere “speculation”) of a material failure by the responding party to meet its obligations.” The Sedona Principles, Third Edition: Best Practices, Recommendations & Principles for Addressing Electronic Document Production, 19 Sedona Conf. J. 1, 123 (2018). At the same time, the Sedona Conference and this district have recognized that search and retrieval parameters and techniques are appropriate topics for discussion at an early meet and confer session. Id. at 83; ESI Guideline 2.02 (stating that it may be useful for the parties to discuss, as part of their Rule 26(f) meet and confer conference “planned methods to identify discoverable ESI and filter out ESI that is not subject to discovery.”).
The Court expects cooperation relating to the search and production of ESI, and “[c]ooperation in reasonably limiting ESI discovery requests on the one hand, and in reasonably responding to ESI discovery requests on the other hand, tends to reduce litigation costs and delay.” ESI Guideline 1.02; see also Comment 3.b., 19 Sedona Conf. J. at 76 (“Cooperation among counsel can enhance the meet and confer process, reduce unnecessary delay and expense associated with non-merit issues, and foster the overriding objectives of Rule 1.”). Pursuant to the Federal Rules of Civil Procedure, the parties are required to confer on and prepare a discovery plan, and their conference must address ESI issues. Fed. R. Civ. P. 26(f)(3)(C). The Northern District's accompanying Checklist for Rule 26(f) meet and confers regarding ESI guides the parties to discuss “search method(s), including specific words or phrases or other methodology, that will be used to identify discoverable ESI and filter out ESI that is not subject to discovery,” as well as the “time period during which discoverable information was most likely to have been created or received” and custodians.[2] The Checklist does not specify “the timing of the discussion of the discussion about these topics, [which] may depend on the nature and complexity of the matter.”
To strike the appropriate balance, the Court modifies Plaintiffs proposed Paragraph 5.b as follows:
At the appropriate time, a Party that intends to use or is using a search methodology (if any) to identify documents for production, shall disclose to the Requesting Party, the search terms, keywords, date limitations, and custodians, and shall specify what custodial and non-custodial sources that the search methodology will or was run against. The custodial sources will presumptively be those identified in the disclosures set forth in Paragraph 5.a. The parties shall discuss search methods at an early meet and confer conference, including the retrieval parameters identified in the preceding sentence, to engage in meaningful cooperation that reduces the costs and risk associated with ESI production.
While the early meet and confer conference may not entail an exhaustive discussion of every search to be deployed, the Requesting Party may, at a later point in the discovery process, request a meet and confer to address the adequacy of a search method or search term, if there is a specific, tangible concern that the Requesting Party has identified to the Producing Party. In such instances, the Producing Party shall provide to the Requesting Party any information that it has not already provided that is necessary to evaluate the effectiveness of the method or the search, such as the search term hit reports pursuant to 5.d.
*5 At the early meet and confer conference where the parties address ESI search methods, the parties shall also discuss what method(s) will be used for documents that are reasonably believed to be responsive and for which text-based search technologies are fundamentally in effective, such as images, spreadsheets, hard copy documents, or certain foreign language documents where the parties do not have suitable search terms in such language. The Producing Party shall disclose to the Requesting Party the specific method(s) it intends to use to search for such documents, and the Requesting Party may make a reasonable request for information that relates to the method(s)' effectiveness.
Should a Producing Party wish to use any method or methodology not specifically addressed by this Order to exclude documents from any review or production, notice to the Receiving Party must be made prior to the use of any such method or methodology. If a Party intends to use Technology Assisted Review (“TAR”) (including Continuous Active Learning), the Parties will agree to a separate TAR protocol to govern that process. If the Parties are unable to reach agreement on a TAR protocol, they will submit any disputed issues to the Court or its designee for resolution.
Plaintiffs' proposed Paragraphs 5.c and 5.d address search terms and hit reports, respectively. In Paragraph 5.c Plaintiffs propose that the parties shall meet and confer to discuss the Producing Party's proposed search terms and methodology, and prior to or during such meet and confers, the Producing Party shall provide all associated information required by the Court. Defendants do not appear to object to meeting and conferring regarding search terms: they have agreed to meet and confer about methods to search ESI; they propose a sentence stating, “The Court encourages the parties to confer on a process to test the efficacy of the search terms[;]” and they agree that the parties may agree jointly to modify limits on the number of search terms that the Court may impose. Defendants appear amenable to meeting and conferring regarding the selection of search terms and to discuss their efficacy. For clarity, the Court adopts Plaintiff's proposed language in Paragraph 5.c requiring that the parties shall meet and confer to discuss the Producing Party's proposed search terms and methodology. The Court also adopts Defendant's proposed language, with certain modifications, concerning the tailoring of search terms.
The parties' proposals diverge with respect to establishing a requirement that the Producing Party provide search term hit reports in certain circumstances, and limiting the number of search terms. Plaintiffs' proposed Paragraph 5.c provides that prior to or during meet and confers addressing search terms, “the Producing Party shall provide all associated information required by this Order.” The proposed Paragraph 5.b states, “At all meet and confers where the topic is adequacy of search terms to identify documents subject to production, the Producing Party shall be required to provide search term hit reports pursuant to paragraph 5.D, below.” Defendants' proposed protocol does not address hit reports.
The effectiveness of meet and confers depends on the parties' candor, and this is especially true when the parties are addressing potential search terms. Defendants are evidently open to engaging in a process to test the efficacy of the search terms, but their proposal does not offer an alternative to Plaintiffs' position that the Producing Party provide hit reports containing certain data points. Accordingly, the Court accepts Plaintiffs' proposal that hit reports be provided as part of the meet and confer process regarding search terms.
*6 Turning to the number of search terms, Defendants propose that each requesting party limit its ESI production requests to five search terms per custodian per party—the same limited imposed by the Patent Model ESI Order. ECF No. 248-3 at 4. Plaintiffs do not propose to limit the number of search terms that a requesting party may use. Defendants, however, have not persuaded the Court that the default limit imposed by the Patent Model ESI Order ought to apply to this case. Instead, the Court sets a limit of fifteen search terms per custodian per Party, which was the limit in the Nazemian ESI Order.
In light of the Court's determinations above, Paragraph 5.c shall read as follows:
Where the Parties agree that potentially responsive ESI shall be searched using search terms, the Parties shall meet and confer to discuss the Producing Party's proposed search terms and methodology. Prior to or during such meet and confers, the Producing Party shall provide all associated information required by this Order. At all meet and confers where the topic is adequacy of search terms to identify documents subject to production, the Producing Party shall provide search term hit reports pursuant to Paragraph 5.d, unless the parties agree otherwise.
Each Requesting Party shall limit its ESI production requests to a total of fifteen search terms per custodian per party. The search terms shall be narrowly tailored to particular issues. Indiscriminate terms, such as the producing company's name or its product name, which are inappropriate unless combined with narrowing search criteria that sufficiently reduce the risk of overproduction. A conjunctive combination of multiple words or phrases (e.g., “computer” and “system”) narrows the search and shall count as a single search term. A disjunctive combination of multiple words or phrases (e.g., “computer” or “system”) broadens the search, and thus each word or phrase shall count as a separate search term unless they are variants of the same word. Use of narrowing search criteria (e.g., “and,” “but not,” “w/x”) is encouraged to limit the production.
If, after disclosure of the Producing Party's proposed search method, search parameters, and search terms, and after a reasonable meet-and-confer process, a Requesting Party has a specific, tangible, evidence based reason to believe that the Producing Party's proposals regarding search, retrieval, and production would result in deficiencies in production, the Requesting Party may object to the proposal and make requests for different or additional search methods, parameters, or search terms. Such requests shall only be made after the Parties have met and conferred as to the alleged deficiencies identified by the Requesting Party. If the issue is not resolved within fourteen (14) calendar days after the objection, either Party may thereafter immediately submit the dispute to the Court or its designee for resolution.
Plaintiffs' proposed Paragraph 5.d is adopted in full.
Plaintiffs' proposed Paragraph 5.e states that the parties “will meet and confer about methods to validate each Producing Party's review process[.]” ECF No. 249-4 at 6. Defendants' proposed ESI protocol does not address validation. It is appropriate for the ESI protocol to meaningfully address validation procedures, and the Court modifies Plaintiff's proposed Paragraph 5.e as follows, modified from In re Social Media Adolescent Addiction, 2024 WL 1786293, at *6-7:
Each Producing Party shall take reasonable steps to validate its own review process (i.e., using quality control measures to determine whether its production is missing relevant ESI or contains substantial amounts of irrelevant ESI) and shall take necessary adjustments to its process. After a Requesting Party reviews a Producing Party's production, the Requesting Party may request additional information regarding the validation method(s) used by the Producing Party and the Producing Party shall disclose the level of end-to-end recall (the percentage of responsive documents in the collection that were identified as responsive by the Producing Party's methodology). If there are remaining disputes after the Producing Party discloses the level of end-to-end recall, the parties shall meet and confer in good faith to resolve any remaining disputes. Such meet and confer efforts shall include a discussion of the tool and parameters used to calculate recall levels.
D. Paragraph 6
*7 Plaintiffs propose adding a paragraph addressing “Enterprise messaging, mobile device data, and other communication tools.” ECF No. 249-4 ¶ 6. The Court recognizes that the parties' messaging platforms used during the course of business (Slack, Microsoft Teams, Telegram, etc.) and mobile devices may contain responsive ESI and accordingly approves the addition of Plaintiffs' Paragraph 6. Id.
E. Paragraph 7 and 8
The Court approves Plaintiffs' proposed edits to Paragraph 7. ECF No. 249-4 ¶ 7. It likewise approves the parties' agreed-upon changes to Plaintiffs' Paragraph 8 (Defendants' Paragraph 7) addressing phasing ESI production. Id. ¶ 8; 248-3 ¶ 7.
F. Paragraph 9
Plaintiffs, at Paragraph 9, and Defendants, at Paragraph 8, propose adding specifications about deduplication. Defendants' proposal includes more permissive language (“the Producing Party shall take reasonable steps to produce only one such identical copy”) while Plaintiffs' proposal is more direct (“the Producing Party shall produce only one such identical copy”) but the proposals are otherwise the same. ECF No. 249-3 ¶ 8. The Court directs the parties to use the language in Defendants' proposed Paragraph 8.
G. Paragraphs 10 and 11
Plaintiffs propose protocols for handling embedded files (Paragraph 10). Defendants object to producing embedded files, asserting that such an obligation would be unduly burdensome. Defendants, however, do not explain why such a production would be unduly burdensome when embedded files are included in ESI production in several other comparable cases. See e.g., Nazemian ESI Order at 7-8; In re Uber Techs., Inc., Passenger Sexual Assault Litig., No. 23-md-03084-CRB-LJC, 2024 WL 1772832, at *6 (N.D. Cal. Apr. 23, 2024); In re Facebook, Inc. Consumer Priv. User Profile Litig., No. 18-md-2843 VC (JSC), 2021 WL 10282215, at *30 (N.D. Cal. Sept. 29, 2021). The Court grants Plaintiffs' request to add Paragraph 10, requiring the production of embedded files.
Plaintiffs propose protocols for handling the production of hyperlinked documents (Paragraph 11). Defendants' arguments misapprehend Plaintiffs' proposal. Defendants claim that Plaintiffs' proposal would require the production of any hyperlinks that appear in any document or communication, along with metadata connecting the document in native format to the document. Plaintiffs, however, appear to acknowledge, to some extent, that technological limitations will restrict which hyperlinks may be “automatically produced,” and propose that upon a reasonable request by the Receiving Party, the Producing Party will produce such hyperlinked documents. ECF No. 249-4 at 8. In addition, Plaintiffs' proposal calls upon the Producing Party to use “available technology” to match hyperlinked documents with email or familial documents, and the production of related metadata, e.g. BEGATTACH and ENDATTACH, would only be required where the technology is available to extract such information. Plaintiffs' proposal does not appear as onerous as Defendants contend, though it leaves out important details, such as how many hyperlinks may be requested by a receiving party.
To enhance the ESI protocol's clarity, the Court adopts, with some modifications, the provisions for hyperlinks, the production of hyperlinked documents, and hyperlink metadata from Nazemian ESI Order at 8-9 (Section 16). Accordingly, Paragraph 11 and its subparagraphs shall read as follows:
*8 11. HYPERLINKS
Except as provided in this paragraph and its subparagraphs, the Producing Party shall not be required to produce documents solely because they are hyperlinked within the Producing Party's production. For purposes of this protocol, “hyperlinked documents” include but are not limited to documents shared via Google Workspace, Microsoft Office's “Share Documents via Link” feature, or similar cloud-based document storage and sharing platforms.
The Parties agree to meet and confer in good faith to resolve any disputes concerning the appropriateness of production of the requested documents and metadata. Nothing in this paragraph and its subparagraphs shall prevent the Producing Party from producing hyperlinked documents that are relevant or responsive to discovery requests in addition to the documents produced by specific request by the Receiving Party.
a. PRODUCTION OF HYPERLINKED DOCUMENTSThe Parties may request the production of the current version or contemporaneous version of a responsive and non-privileged document or communication that is within the Producing Party's possession, custody, or control and is hyperlinked within the Producing Party's production by providing to the Producing Party a list of hyperlinks and corresponding Bates numbers in which the hyperlinked document is referenced. If the hyperlinked document has already been produced, the Producing Party may alternatively provide the Receiving Party with the document's Bates number. Each Party (with Named Plaintiffs constituting a single Party, and Stability AI entities constituting a single Party) may make up to two hundred such requests. The parties agree to meet and confer in good faith to resolve any disputes concerning the appropriateness of the production of the requested documents and communications and any reasonable request to increase the limit on the total number of hyperlinked documents and communications to be produced. If a Requesting Party believes that more requests are warranted and the Receiving Party has not agreed to produce additional hyperlinked documents, the Requesting Party may initiate the joint letter-brief process pursuant to the Court's dispute resolution procedures set forth in Standing Order, Section F.5.
b. PRODUCTION OF HYPERLINK METADATAThe Parties acknowledge that the current version of the hyperlinked document (i.e., the version produced pursuant to the paragraphs concerning hyperlinks) may differ from the version of the document at the time the hyperlink was sent. The Parties also acknowledge that the metadata of such document(s) may not reflect the metadata for requested documents at the time the documents were hyperlinked. For hyperlinked documents produced pursuant to these paragraphs concerning hyperlinks, the Producing Party will use available technology to match hyperlinked documents with the email or familial document to which the documents were attached in Native format, and the Producing Party will use available technology to provide the BEGATTACH and ENDATTACH metadata fields to show the family relationship.
H. Paragraph 12
*9 At Paragraph 12, Plaintiffs propose adding specifications about how a producing party must collect and produce structured data, including requirements that the producing party inform the receiving party about all possible formats for the structured data, and, if parties cannot agree on the method of production, the producing party “make the information available for inspection on this system where it is currently maintained, or otherwise provide direct access to the data.” ECF No. 249-4 ¶ 12. Defendants argue that such process could provide the receiving party “unfettered access to their IT infrastructure with no specific showing of burden, need, or proportionality[.]” ECF No. 248 at 5. Plaintiffs recognize that “it is important for the parties to discuss at the outset how [structured data] should be handled.” ECF No. 249-1 (Saveri Decl.) ¶ 12. It is unclear if the parties have engaged in any meaningful discussion about the production of structured data. The parties are therefore ordered to meet and confer to develop a stipulated protocol for producing and reviewing structured data.
I. Paragraph 13
The Court accepts Plaintiffs' addition of Paragraph 13 to the extent that Defendants have not proposed deleting certain text. ECF Nos. 249-4 ¶ 4; 248-3 ¶ 9. The examples of how Model Data may be produced, set forth in Plaintiffs' proposed Paragraph 13, do not obligate Defendants to produce Model Data in those formats and the inclusion of this sentence do not waive any arguments Defendants might wish to make about producing Model Data. The parties may discuss those examples during their meet and confer conference regarding model data. Paragraph 13 shall read as follows:
To the extent such data is not addressed in this protocol, the Parties will meet and confer to address the production of data related to the Models at issue. Given the size, structure and format of such data, the Parties agree to discuss how Model Data, if any, should be produced.
J. Paragraphs 14 and 15
Plaintiffs propose adding detailed specifications about what information must be included in privilege logs and a procedure for challenging asserted privileges. ECF No. 249-4 ¶¶ 14-15. The Court agrees that not establishing clear requirements for privilege logs and privilege log disputes at the outset of discovery will likely invite future disputes. Defendants have objected to Plaintiffs' proposed specifications for the privilege logs, but these requirements are in line with privilege log requirements in other cases, see e.g., Nazemian ESI Order at 9-11; In re Uber Techs., Inc., Passenger Sexual Assault Litig., No. 23-md-03084-CRB-LJC, at ECF No. 396 (¶¶ 14-17), and Defendants fail to cite to any specific authority that discourages against these provisions. The Court accepts Plaintiffs' proposal for privilege logs and the procedure for challenging asserted privileges at Paragraphs 14 and 15, respectively.
K. Paragraph 16
The parties agree regarding “DOCUMENTS PROTECTED FROM DISCOVERY”. See Plaintiffs Paragraph 16, ECF No. 249-4 at 12; Defendants' Paragraph 11, ECF No. 248-3 at 6.
L. Appendices
Plaintiffs have proposed two appendices: Appendix A, which sets forth technical specifications for the production of ESI, and Appendix B, which addresses metadata fields. Defendants object to both, though they only provide a counterproposal for metadata fields. Defendants make only a conclusory argument that Plaintiffs' proposal for technical specifications for ESI production is onerous. Therefore, the Court adopts Plaintiffs' proposed Appendix A, ECF No. 249-4 at 15-17.
With regard to the appendix for metadata fields, given Defendants' representation that complying with Plaintiffs' proposal would require manual entry, be “technically onerous[,]” and likely to result in error, the Court directs the parties to use Defendants' metadata fields. ECF Nos. 248 at 5; 248-3 at 9-11. This ruling is made without prejudice and Plaintiffs may renew their request for the additional fields upon showing that their need outweighs the burden on Defendants.
III. PROTECTIVE ORDER
The parties' only outstanding dispute regarding their protective order is the definition of “source code.” See ECF No. 248-5 at 4. Plaintiffs define source code as “the human readable set of instructions written in a programming language that defines the operations of a software application, including the algorithms, logic, and functions that shape the program's behavior.” ECF No. 249-3 § 2.18. Defendants seek to use the definition provided in the Patent Model Protective Order, which defines source code as “computer code and associated comments and revisions histories, formulas, engineering specifications, or schematics that define or otherwise describe in detail the algorithms or structure of software or hardware designs, disclosure of which to another Party or Non-Party would create a substantial risk of serious harm that could not be avoided by less restrictive means.” ECF No. 248-2 § 2.9; Model Protective Order for Litigation Involving Patents, Highly Sensitive Confidential Information and/or Trade Secrets, U.S. District Court, Northern District of California, https://www.cand.uscourts.gov/forms/model-protective-orders/ (last visited Jan. 22, 2025). Although the parties are not required to use the Patent Model Protective Order definition, the Court agrees with Defendants that using the broader definition provided in the Patent Model Protective Order is appropriate here. See In re Meta Pixel Healthcare Litigation, No. 22-cv-03580, 2023 WL 3483296, at *1 (N.D. Cal. May 16, 2023) (adopting a definition of “source code” that “generally tracks” the language in the Patent Model Protective Order). The Court recognizes that “comments and revisions histories, formulas, engineering specifications, or schematics” that define or describe code may be as sensitive—and thus warrant the same safeguards—as the code itself. ECF No. 248-2 § 2.9. The parties are ordered use the Patent Model Protective Order's definition of source code.
IV. DEPOSITION PROTOCOL
*10 The parties submitted an additional dispute regarding their deposition protocol. ECF No. 257. Plaintiffs request to conduct up to 60 depositions of Defendants' fact witnesses. Defendants object, countering that each side should be entitled to take up to twelve depositions.
Federal Rule of Civil Procedure 30(a) provides that a party may take up to ten depositions without leave of the court. Courts “may alter the limits ... on the number of depositions ...or on the length of depositions under Rule 30.” Fed. R. Civ. P. 26(b)(2)A). In determining whether to permit more than ten depositions, the court must consider “whether: (i) the discovery sought is unreasonably cumulative or duplicative, or is obtainable from some other source that is more convenient, less burdensome, or less expensive; (ii) the party seeking discovery has had ample opportunity by discovery in the action to obtain the information sought; or (iii) the burden or expense of the proposed discovery outweighs its likely benefit[.]” Authentec, Inc. v. Atrua Techs., Inc., No. C 08-1423 PJH, 2008 WL 5120767, at *1 (N.D. Cal. Dec. 4, 2008). Although “courts will generally not grant leave to expand the number of depositions until the moving party has exhausted the ten depositions permitted as of right” (id.), the Northern District regularly departs “from this rule where the complexity of the case clearly warranted more than ten depositions.” Sayce v. Forescout Techs., Inc., No. 20-cv-00076, 2024 WL 3579002, at *1 (N.D. Cal. July 11, 2024); see Del Campo v. American Corrective Counseling Servs., Inc., No. C-01, 2007 WL 3306496 (N.D. Cal. Nov. 6, 2007); Kadrey v. Meta Platforms, Inc., No. 23-cv-03417, 2024 WL 4502099, at *2 (N.D. Cal. Oct. 15, 2024) (permitting Plaintiff to take twenty-five fact depositions in copyright dispute regarding generative AI). “A party seeking to exceed the presumptive number of depositions must make a particularized showing of the need for the additional discovery.” Authentec, 2008 WL 5120767, at *1.
Plaintiffs argue they are entitled to sixty depositions because of “the number of corporate defendants, the technical nature of the infringement, and the vast scale of the alleged misconduct.” ECF No. 257 at 1. They reason that taking sixty depositions, or fifteen per defendant, is reasonable given the scope of the case. Id. at 2. This will permit them to take twenty to twenty-five depositions of “core technical witnesses”, fifteen to twenty depositions of witnesses in corporate leadership roles (including “executive leadership, project managers, and strategic decision-makers”), and fifteen to twenty depositions of Defendants' “business operations personnel.” Id. at 2-3. Defendants object, arguing that Plaintiffs' request “lacks any particularized justification” and that Plaintiffs' “proposed breakdown of total deponents...is arbitrary and untethered to any document, filing, or fact specific to any Defendant.” Id. at 4. They contend that most of the Defendant companies are “early-stage, lean startups” with relatively few employees, and permitting Plaintiffs to depose “extraneous employees” would result in duplicative testimony and pose a significant burden on Defendants. Id. at 4-5.
The Court agrees with Plaintiffs' overall position that exceeding the ten-deposition limit is warranted here given the “complexity of the case[,]” but is mindful of the potential burden on Defendants and the risk that allowing Plaintiffs sixty depositions would be unreasonably cumulative or duplicative. Sayce, 2024 WL 3579002, at *1; see Fed. R. Civ. P. 26(b). The Court finds it appropriate at this time to grant Plaintiffs leave to take thirty depositions, to be allocated at Plaintiffs' discretion. Plaintiffs may renew their request for additional depositions if they determine they are necessary after taking the first thirty depositions. The thirty-deposition limit does not include depositions of non-parties, expert witnesses, of former employees. This limit does not change the parties' agreement regarding 30(b)(6) depositions at ¶ 22 of the Stipulated Order Entering Deposition Protocol. ECF No. 257-1 ¶ 22.
V. CONCLUSION
*11 Having resolved the disputes discussed above concerning the terms of the ESI protocol, the parties are ordered to incorporate the Court's rulings from this Order into a final Stipulated and [Proposed] ESI Protocol, to be filed with the Court by March 31, 2025. Before March 31, 2025, the parties shall also meet and confer regarding the outstanding issues related to structured data and model data. If the parties are able to reach an agreement to resolve these issues, they shall include the necessary provisions in the final Stipulated and [Proposed] ESI Protocol. If parties are unable to reach an agreement they shall file by April 7, 2025 a joint discovery letter concerning their respective proposals, following the procedure in Section F.5 of the undersigned's Standing Order.
The parties shall file a Stipulated Protective Order, consistent with this Order, by March 31, 2025.
IT IS SO ORDERED.