Opinion
22-cv-02366-RS (RMI)
03-14-2023
SHARED PARTNERSHIP, Plaintiff, v. META PLATFORMS, INC., Defendant.
ORDER RE: DISCOVERY DISPUTES RE: DKT. NOS. 47, 49
ROBERT M. ILLMAN UNITED STATES MAGISTRATE JUDGE
Now pending before the court are a pair of discovery dispute letter briefs (dkts. 47, 49) that, generally speaking, represent Plaintiff's request to compel the production of certain discovery. It appears that the two letters are substantially the same, with the exception that the first letter appears to have been rushed to court before Defendant could articulate its position as to one of the issues raised by Plaintiff (see dkt. 47 at 3-4; see also dkt. 49 at 1, 4-5). Defendant submits that Plaintiff has involved the court prematurely, that the dispute is not quite ripe for a motion to compel, and that the parties should continue to engage one another in the discovery process. See Ltr. Br. (dkt. 49) at 2. For the reasons stated below, the court wholeheartedly agrees.
Plaintiff has presented two issues. In the first, Plaintiff complains of the timing of Defendant's document production. See id. at 1-2. In this regard, Plaintiff submits that it tendered a set of interrogatories and requests for production in late October of 2022, and that Defendant's responses - which were served in mid-December of 2022, did not contain substantive information but were rather limited to objections (presumably as to some requests) and statements to the effect that responsive documents and information would be produced on a rolling basis (again, presumably as to certain other requests). Id. at 1. Based only on the argument that Plaintiff should not be forced to agree to an speculative extension of the fact discovery deadline (currently set for September 21, 2023) and risk pushing back an already-scheduled trial date of April 2024, Plaintiff submits that Defendant should now be forced “to complete its entire document production by April 7, 2023” because it “has had more than enough time - and has more than enough resources -to accomplish this.” Id. at 2.
Defendant responds to the effect that Plaintiff's 41 document requests are relatively broad and that they seek, among other things, all documents relating to Plaintiff, all documents regarding Plaintiff's rejected advertisements on Defendant's platform, and all documents concerning Defendant's use of artificial intelligence to reject any and all advertisements in general (that is, not simply information limited to the rejection of Plaintiff's own advertisements - the subject of this action). Id. at 2. Nevertheless, Defendant submits that through its December responses, it agreed to search for and produce several categories of information including documents relating to Plaintiff's rejected advertisements, notifications regarding those advertisements, and communications between the Parties. Id. Following discussions regarding the scope, relevance, and proportionality of the requested discovery, Defendant states that it has agreed to produce a narrowed set of documents for the majority of Plaintiff's requests, and that it is working diligently to gather responsive, non-privileged documents for production. Id. at 2-3.
However, Defendant contends that there are three remining hurdles. As to the first, Defendant states that it “cannot produce any documents until a stipulated protective order is entered between the parties and Court, and the parties are currently negotiating the substance of the protective order so that [Defendant] can begin producing documents.” Id. at 3. The court finds that this hurdle is easy to surmount. The Parties are herewith ORDERED to promptly meet and confer in a good faith effort to iron out their differences as to their respective views regarding the substance of the protective order necessary such that Defendant can begin producing documents. The court recommends, but does not require, using this district's model protective order at least as a starting point. In the event that the Parties cannot arrive at a stipulated resolution, the Parties are FURTHER ORDERED to clearly and concisely present what disputes remain in this regard through a jointly-filed letter brief conforming to the requirements set forth in ¶ 4 of the undersigned's general standing order.
See https://www.cand.uscourts.gov/wp-content/uploads/judges/illman-rmi/RMI-Standing-Order-Revised9-16-22.pdf
As to the second hurdle, Defendant submits that it “continues to work diligently to search for documents responsive to the majority of [the] requests it has agreed to produce” and that it “has already identified tens of thousands of potentially responsive documents to [Plaintiff's] rejected advertisements and notifications regarding those rejected advertisements.” Id. at 3. Defendant explains that the responsive discovery is voluminous and time-consuming to gather and produce due to the fact that Plaintiff's advertising relationship on Defendant's platform “goes back over ten years, and the volume of data is substantial,” as well as the fact that “the process to search for, review, and pull those documents, which began months ago, is manual - [and] there are thus multiple, time-consuming steps [Defendant] must undertake before it can produce information regarding [Plaintiff's] disapproved advertisements, of which there were many.” Id. Defendant then adds that this is why it has focused on only potentially responsive documents - that is, those related to Plaintiff's rejected advertisements - rather than documents related to Defendant's more “general purported use of artificial intelligence.” Id. At bottom, Defendant contends that “[t]hese documents are probative of whether [Plaintiff's] ads were rejected and what notice [Plaintiff] received regarding those ads, which is directly relevant to [the] fraud-based claims.” Id. The court finds that the assertion of this hurdle is closely intertwined with the second issue raised by Plaintiff's request to compel; accordingly, the court will address it below.
As to the third hurdle, Defendant submits that it has not yet provided a specific date by which it will complete production of all responsive, non-privileged documents because it “continues to search for responsive documents, the extent of those documents, and [because it is currently engaged in] the labor-intensive process of pulling those potentially responsive documents.” Id. In short, Defendant submits that it “has been working for months to gather and review documents for production once the protective order is entered,” and that it “intends on producing documents on a rolling basis shortly thereafter.” Id.
In essence, this dispute boils down to Plaintiff's desire to rush Defendant through the discovery process based only on a thin justification that Defendant “has had more than enough time - and has more than enough resources - to accomplish this.” Id. at 2. The court finds this notion to be unpersuasive. The alleged magnitude of Defendant's resources, and the passage of a few months of time does not justify arbitrarily setting a production deadline of April 7, 2023, as Plaintiff suggests. Once the Parties' dispute about the terms of the forthcoming protective order (either by stipulation or by court order) are resolved, the court sees no reason - and Plaintiff has certainly not suggested one - that Defendant's production should not proceed on the rolling basis it has suggested and justified. Accordingly, Plaintiff's request to force Defendant to complete its production by April 7, 2023, is DENIED.
As to the second issue, Plaintiff seeks an order “making clear that [Defendant's] use of artificial intelligence to reject ads-including what it did and did not know about the ability of artificial intelligence to provide explanations for ad rejections-is both relevant and discoverable.” Id. Plaintiff then adds that “[t]his dispute is implicated by the following discovery requests: a) Interrogatories 2-4 and b) RFP Nos. 15-20; 25-34; and 38-41.” Id. at 4. Plaintiff has not included those discovery requests as attachments or exhibits to its letter brief. However, this omission is unimportant for two reasons.
First, Plaintiff has failed to convince the undersigned as to the probative value of this overwhelmingly broad request. Plaintiff points to the fact that its fraud claim survived dismissal when the presiding judge found that because Plaintiff had asserted that Defendant's advertising policies contained a fraudulent statement to the effect that advertising partners would be provided with sufficient details to bring their rejected ads into conformity with Defendant's policies, it was therefore plausible that a reasonable consumer would concur with Plaintiff's interpretation of the policies, notwithstanding Defendant's contrary interpretation. Id. In short, in the context of deciding a motion to dismiss, the presiding judge found that it was plausible that Defendant knew or should have known that it could not comply with this expectation due to its averred reliance on artificial intelligence. Id. However, simply because an allegation has been found to be sufficiently plausible such that its associated claim survives dismissal, it does not necessarily follow that discovery thereunder will automatically be found to be relevant to that claim at the discovery phase of a case. See e.g., Elgindy v. AGA Serv. Co., No. 20-cv-06304-JST (RMI), 2021 U.S. Dist. LEXIS 211772, at *12-13 (N.D. Cal. Nov. 2, 2021) (“. . . the scope of Plaintiffs' 4 RFPs are more geared to proving or disproving Plaintiffs' [own] speculative and editorial musings rather than the[] actual claims. [R]elevance must be tethered to a claim, not simply to an allegation as seems to be implied by Plaintiffs' suggestion . . .”). In this regard, the court finds - based on Plaintiff's presentation (limited to the plausibility finding at the motion to dismiss stage) - that generalized information about Defendant's suggested use of artificial intelligence to reject advertisements other than Plaintiff's own advertisements would have very little - if any - probative value to Plaintiff's claims.
Second, assuming arguendo that the relevance of this request had been established, Plaintiff has expended no effort in justifying the proportionality of such a broad request, while Defendant has made a convincing case about its disproportionality. See id. at 3-4, 5-6. It should be noted here that in order to succeed on a motion to compel, a moving party bears the burden of not only demonstrating relevance, but also demonstrating that it has satisfied proportionality and other requirements of Rule 26. See e.g., Rodriguez v. Barrita, Inc., No. 09-04057 RS (PSG), 2011 U.S. Dist. LEXIS 134079, at *4 (N.D. Cal. Nov. 21, 2011). In light of this, courts are required to limit discovery if its burden or expense outweighs its likely benefit; this is “the essence of proportionality,” a frequently ignored or overlooked discovery principle. See e.g., Hegarty v. Transamerica Life Ins. Co., No. 19-cv-06006-MMC (RMI), 2021 U.S. Dist. LEXIS 203386, at *5 (N.D. Cal. Oct. 21, 2021); see also Williams v. Condensed Curriculum Int'l, No. 20-cv-05292-YGR (RMI), 2021 U.S. Dist. LEXIS 211782, at *8-9 (N.D. Cal. Nov. 2, 2021); see also Mediatek, Inc. v. Freescale Semiconductor, Inc., No. 11-5341 YGR (JSC), 2013 U.S. Dist. LEXIS 122911, at *18-19 (N.D. Cal. Aug. 28, 2013) (“[U]nder Rule 26's proportionality analysis, the Court must weigh the burden versus the benefit of the discovery sought . . . [and] strike a balance between ensuring that parties have access to relevant information regarding their claims and defenses, and the burden on the producing party of providing the discovery sought.”).
As stated by Defendant, despite the fact that Plaintiff's artificial intelligence theory was found plausible at the motion to dismiss stage, by no means does it logically follow (as Plaintiff's arguments imply) that broad-based and wholesale discovery into all aspects of Defendant's purported use of artificial intelligence to reject the advertisements of all of its many customers would be proportional to the needs of Plaintiff's case which involves a drastically smaller subset of events. Accordingly, Plaintiff's request for a finding as to the general discoverability of Defendant's use of artificial intelligence to reject advertisements other than its own is DENIED.
In closing, the court also feels compelled to admonish the Parties to work together more efficiently going forward to resolve their discovery disputes with minimal need for court intervention. “[T]this court is not a clearinghouse for resolving insignificant, unripe, or moot discovery disputes. Instead, counsel should be mindful that discovery is the responsibility of the parties and turning to the court to resolve a discovery dispute should only be used as a last resort.” Miranda Dairy v. Harry Shelton Livestock, LLC, No. 18-cv-06357-RMI, 2020 U.S. Dist. LEXIS 197775, at *13 (N.D. Cal. Oct. 22, 2020) (“While the court does have a duty to actively and effectively manage the discovery process such as to ensure that the parties have efficient access to necessary discovery while eliminating unnecessary and wasteful discovery, at the same time, it goes without saying that this duty is not automatically triggered each and every time counsel can conjure up one or another thing about which to disagree. Attorneys practicing in this court are expected to seek court involvement in discovery disputes only as a last resort, and only when the dispute implicates truly significant interests that counsel cannot resolve through reasonable cooperation during the meet-and-confer process. Neither party has presented any indication at all that this was done here.”); see also Olesczuk v. Citizens One Home Loans, No. 2:16-cv-01008-GMN-NJK, 2016 U.S. Dist. LEXIS 153342, at *5 (D. Nev. Nov. 4, 2016) (same); Morgan Hill Concerned Parents Ass'n v. Cal. Dep't of Educ., No. 2:11-cv-3471 KJM AC, 2016 U.S. Dist. LEXIS 109451, at *7 (E.D. Cal. Aug. 17, 2016) (same); McLain v. Daniel N. Gordon, PC, No. C09-5362BHS, 2010 U.S. Dist. LEXIS 154842, at *3 (W.D. Wash. July 28, 2010) (same).
IT IS SO ORDERED.