Regarding Docket No. 147

On February 24, 2022, the parties filed a joint discovery letter brief regarding their instant discovery dispute. (Dkt. 147.) On March 8, 2022, this matter was referred to the Undersigned for Discovery. (Dkt. 153.)

Plaintiff UAB “Planner5D” alleges that Defendants Princeton University, Meta Platforms, Inc., and Facebook Technologies LLC misappropriated Plaintiff’s trade secrets and infringed Plaintiff’s copyrights. (Dkt. Nos. 53, 90. 112.) Specifically, Plaintiff alleges that Defendant Princeton “scraped” Plaintiff’s works and then converted them into a “computer vision tool called the SUNCF dataset” and that Defendant Meta used that tool to develop products and services in the “scene-recognition market” – a “key component” of the “metaverse.” (Dkt. No. 147.) Plaintiff seeks information regarding damages in the form of four requests: one interrogatory and three requests for production of documents. (Dkt. Nos. 147-1, 147-2.)

Interrogatory No. 14 states:

Describe every OBJECT/SCENE MACHINE LEARNING PROJECT that YOU conducted, sponsored, or participated in that used or uses an INTERIOR OBJECT/SCENE DATASET, by identifying the (i) the PERSON(s) primarily responsible for the Project, (ii) the machine learning algorithm(s) used or developed in that Project, (iii) the INTERIOR OBJECT/SCENE DATASET(s) used in that Project, and (iv) all internal or external research papers referencing that Project.

(Dkt. No. 147-1.) After objecting, Defendant Meta provided a lengthy response identifying five “machine learning research projects.” (Dkt. No. 147-1.)

In Request for Production of Documents No. 27, Plaintiff seeks: “ALL RECORDS constituting, referencing, or concerning any evaluation of characteristics and/or strengths or weaknesses of any INTERIOR OBJECT/SCENE DATASET.” (Dkt. No. 147-2.) Defendant Meta objected to this request and responded that it would produce “nonprivileged, relevant documents about the evaluation of characteristics, strengths or weaknesses of the SUNCG Dataset, to the extent such documents exist and are located after a reasonably diligent search. In an amended response, Defendant Meta further agreed to produce “nonprivileged, relevant documents about the projects identified in Meta’s response to Interrogatory No. 14[.]” (Dkt. No. 147-2.)

In Request for Production of Documents No. 28, Plaintiff seeks: “ALL RECORDS constituting, referencing, or reflecting any projection of the value (monetary or otherwise, and whether quantitative or qualitative) of any INTERIOR OBJECT/SCENE DATASET.” (Dkt. No. 147-2.) In response, Defendant Meta objected and agreed to produce “nonprivileged, relevant documents about the projection of the value of the SUNCG Dataset, to the extent such documents exist and are located after a reasonably diligent search.” In an amended response, Defendant Meta further agreed to produce “nonprivileged, relevant documents about the projects identified in Meta’s response to Interrogatory No. 14[.]”

In Request for Production of Documents No. 29, Plaintiff seeks: “A copy of every INTERIOR OBJECT/SCENE DATASET(s) YOU have used for an OBJECT/SCENE MACHINE LEARNING PROJECT.” In its amended response, Defendant Meta objected and agreed to produce “non-privileged, responsive documents sufficient to show where the interior scene or object datasets that were used in the projects listed in Interrogatory No. 14 before this litigation was filed are available for download[.]” (Dkt. No. 147-2.)

Defendant Meta argues that the requests are not relevant to the case in seeking information about its other work in the area of scene recognition using interior object/scene datasets, as Defendant Meta asserts that the sole isuse in the case is whether it used Plaintiff’s intellectual property in a manner that violates the law. Plaintiff argues that the information is relevant to the issue of damages because one measure of damages is the amount “a willing buyer would have been reasonably required to pay a willing seller.” See Oracle Corp. v. SAP AG, 765 F.3d 1081, 1087 (9th Cir. 2014) (discussing hypothetical license for copyright); Cal. Civ. Code §3426.3 (reasonable royalty for trade secret). Plaintiff thus seeks information about Defendant Meta’s willingness to pay for datasets to show the value (or lack thereof) of Plaintiff’s datasets. The Court finds that the requested information is relevant to damages.

Defendant Meta also argues that the requests are not proportional to the case because it must interview hundreds of people to in response and collect and process tens of thousands of documents. Although the Court recognizes that the burden is great, Defendant Meta does not provide any context for proportionality. In assessing proportionality, a court must balance the burden with the scope and needs of the case. Here, Defendant Meta does not meet its burden to show that the balance tips in its favor for these requests.

The Court ORDERS Defendant Meta to respond to the requests. The Court ORDERS the parties to meet and confer regarding a schedule for production.

IT IS SO ORDERED.

Dated: March 18, 2022