Case 3:22-cv-02022 Document 1 Filed 08/30/22 Page 1 of 28 Page ID #1
`
`
`
`IN THE UNITED STATES DISTRICT COURT FOR THE
`SOUTHERN DISTRICT OF ILLINOIS
`
`
`BRAD MARSCHKE, Individually and on
`Behalf of All Others Similarly Situated,
`
`Plaintiff,
`
`vs.
`
`YOUTUBE, LLC and GOOGLE LLC,
`
`Defendants.
`
`
`
`
`
`Case No. 3:22-cv-2022
`
`
`)
`)
`)
`)
`)
`)
`)
`)
`)
`)
`)
`
`
`
`
`
`
`
`CLASS ACTION COMPLAINT AND DEMAND FOR JURY TRIAL
`
`
`
`

`

`Case 3:22-cv-02022 Document 1 Filed 08/30/22 Page 2 of 28 Page ID #2
`
`
`
`Plaintiff Brad Marschke brings this Class Action Complaint against Defendants YouTube,
`
`LLC (“YouTube”) and Google LLC (“Google”) (collectively, “Defendants”) to put a stop to
`
`Defendants’ surreptitious collection, use, and storage of Plaintiff’s and the proposed Class’
`
`(defined below) sensitive biometric identifiers1 and biometric information2 (collectively,
`
`“biometrics”) without obtaining informed written consent or providing the data retention and
`
`destruction policies to consumers. Plaintiff alleges as follows upon personal knowledge as to
`
`himself and his own acts and experiences, and, as to all other matters, upon information and belief,
`
`including investigation conducted by his attorneys.
`
`NATURE OF THE ACTION
`
`1.
`
`Google is one of the largest corporations in the world. Its business focuses on,
`
`among other things, artificial intelligence, Internet search engine technology, online advertising,
`
`cloud computing, computer software, quantum computing, e-commerce, and consumer
`
`electronics.
`
`2.
`
`Google has been referred to as the “most powerful company in the world” and one
`
`of the world’s most valuable brands due to its market dominance, data collection, and
`
`technological advantages in the area of artificial intelligence. It is considered one of the “Big Five”
`
`American information technology companies, alongside Amazon.com, Inc., Apple Inc., Meta
`
`Platforms, Inc. (f/k/a Facebook), and Microsoft Corporation.
`
`3.
`
`Google offers a multitude of products and services beyond its ubiquitous Google
`
`Search, many of which hold dominant market positions, including video sharing through YouTube.
`
`
`1
`“‘Biometric identifier’ means a retina or iris scan, fingerprint, voiceprint, or scan of hand or face
`geometry.” 740 ILCS 14/10.
`
`2
`“‘Biometric information’ means any information, regardless of how it is captured, converted, stored,
`or shared, based on an individual’s biometric identifier used to identify an individual.” Id.
`
`- 1 -
`
`

`

`Case 3:22-cv-02022 Document 1 Filed 08/30/22 Page 3 of 28 Page ID #3
`
`
`
`4.
`
`Google purchased YouTube on November 13, 2006, for $1.65 billion in Google
`
`stock. YouTube is now a wholly owned subsidiary of Google.
`
`5.
`
`Google and YouTube operate the two most visited websites worldwide, google.com
`
`and youtube.com.3
`
`6.
`
`Google.com and youtube.com are also the two most visited websites in the United
`
`States.4
`
`7.
`
`Founded in 2005, YouTube is the largest social media video-sharing platform in
`
`the world.
`
`8.
`
`YouTube’s platform has more than 2.5 billion monthly users who collectively
`
`watch more than one billion hours of videos each day.
`
`9.
`
`As of May 2019, videos were being uploaded to YouTube at a rate of more than
`
`500 hours of content per minute.
`
`10.
`
`YouTube users can use the YouTube platform to, among other things, upload and
`
`share videos, either privately or with the general public.
`
`11.
`
`Once a user uploads a video on YouTube, the user can use various YouTube
`
`features or tools to, among other things, blur out faces within an uploaded video and/or create
`
`thumbnail photographs of various points within a particular video.
`
`12.
`
`As alleged in more detail below, Defendants’ “Face Blur” tool allows a user to
`
`“select the faces” in the user’s particular video that they would “like to blur,” which when applied
`
`
`3 Most Visited Websites by Traffic in the world for all categories, July 2022, SEMRUSH,
`https://www.semrush.com/website/top/.
`
`4 Most Visited Websites by Traffic in United States for all categories, July 2022, SEMRUSH,
`https://www.semrush.com/website/top/united-states/all/.
`
`- 2 -
`
`

`

`Case 3:22-cv-02022 Document 1 Filed 08/30/22 Page 4 of 28 Page ID #4
`
`
`
`and saved, will result in those faces appearing blurry and ostensibly unrecognizable to any viewer
`
`of the video.
`
`13.
`
`In order for a YouTube user to use the “Face Blur” tool, the user must use YouTube
`
`Studio, which Defendants claim is “the home for creators[,]” and allows YouTube users to
`
`“manage [their] presence, grow [their] channel, interact with [their] audience, and make money all
`
`in one place.”5
`
`14.
`
`Upon information and belief based on the investigation of Plaintiff’s counsel,
`
`Defendants’ “Face Blur” tool captures and stores biometric identifiers or information in the form
`
`of scans of face geometry that the tool is used on. However, Defendants do not provide notice or
`
`obtain legally mandated consent from the individual’s whose biometric identifiers or information
`
`is captured, in violation of Illinois’ Biometric Information Privacy Act, 740 ILCS 14/1, et seq.
`
`(“BIPA”). Nor do they post a publicly available retention schedule and guidelines for permanently
`
`destroying the biometric identifiers of Plaintiffs and the Class, as mandated by BIPA.
`
`15.
`
`In addition, Defendants offer YouTube users an opportunity to choose specific
`
`thumbnail pictures culled from an uploaded video. Specifically, YouTube contains a feature that
`
`allows video creators to choose thumbnails for their videos that are auto generated by YouTube.
`
`16.
`
`To be sure, thumbnails with faces, especially faces with more expression, generate
`
`more clicks and views and, as such, Defendants are incentivized to auto-generate thumbnails that
`
`contain faces – especially faces that contain more expression.
`
`17.
`
`Upon information and belief based on the investigation of Plaintiff’s counsel,
`
`Defendants’ thumbnail feature appears to work by scanning the uploaded video and all faces within
`
`the video to identify facial expressions. The purpose of this is to attract the most clicks and views
`
`
`5 Navigate YouTube Studio, GOOGLE, https://support.google.com/youtube/answer/7548152?hl=en (last
`visited Aug. 29, 2022).
`
`- 3 -
`
`

`

`Case 3:22-cv-02022 Document 1 Filed 08/30/22 Page 5 of 28 Page ID #5
`
`
`
`for the uploaded videos. In doing so, Defendants capture and store biometric identifiers or
`
`information in the form of faceprints without providing notice or obtaining legally mandated
`
`consent from the individual’s whose biometric identifiers or information is captured, in violation
`
`of BIPA. Nor do they a post publicly available retention schedule and guidelines for permanently
`
`destroying the biometric identifiers of Plaintiff and the Class, as mandated by BIPA.
`
`18.
`
`Through these practices, Defendants not only disregard their users’ privacy rights,
`
`they also violate BIPA, which was specifically designed to protect Illinois consumers from
`
`practices like Defendants’. In particular, Defendants violated (and continue to violate) BIPA
`
`because they did not:
`
`(a)
`
`properly inform Plaintiff or the Class in writing that their biometric
`
`identifiers were being collected or stored;
`
`(b)
`
`properly inform Plaintiff or the Class in writing of the specific purpose and
`
`length of time for which their biometric identifiers were being collected, stored, and used;
`
`(c)
`
`provide a publicly available retention schedule and guidelines for
`
`permanently destroying the biometric identifiers of Plaintiff and the Class; and
`
`(d)
`
`receive a written release from Plaintiff or the Class to collect, capture, or
`
`otherwise obtain their biometric identifiers.
`
`19.
`
`Accordingly, this Complaint seeks a final judgment: (a) declaring that Defendants’
`
`conduct violates BIPA; (b) requiring Defendants to cease the unlawful activities discussed herein;
`
`and (c) awarding statutory damages to Plaintiff and the Class.
`
`PARTIES
`
`20.
`
`Plaintiff Brad Marschke is a natural person and resident and citizen of the State of
`
`Illinois residing in the County of Effingham. Plaintiff has been a customer of Google and a
`
`registered and active user of YouTube since at least 2008. Within the applicable statute of
`
`- 4 -
`
`

`

`Case 3:22-cv-02022 Document 1 Filed 08/30/22 Page 6 of 28 Page ID #6
`
`
`
`limitations period, Marschke uploaded multiple videos to YouTube and used Defendants’ “Face
`
`Blur” tool or thumbnail creator containing at least one representation of his face, which Defendants
`
`in violation of BIPA captured, scanned for face geometry, and stored without providing Plaintiff
`
`any notice or receiving a written release required by BIPA.
`
`21.
`
`Defendant YouTube, LLC is a limited liability company organized and existing
`
`under the laws of the State of Delaware, and is headquartered in San Bruno, California. YouTube
`
`is a wholly owned subsidiary of Google.
`
`22.
`
`Defendant Google LLC is a limited liability company organized and existing under
`
`the laws of the State of Delaware, and is headquartered in Mountain View, California. Google is
`
`an online advertising technology company providing internet-related products, including
`
`YouTube. Google is owned by Alphabet Inc., a publicly traded company incorporated and existing
`
`under the laws of the State of Delaware and headquartered in Mountain View, California.
`
`JURISDICTION AND VENUE
`
`23.
`
`This Court has jurisdiction pursuant to 28 U.S.C. §1332(d)(2)(A), because at least
`
`one member of the Class is a citizen of a different state than Defendants, there are more than 100
`
`members of the Class, and the aggregate amount in controversy exceeds $5,000,000, exclusive of
`
`interest and costs.
`
`24.
`
`This Court has personal jurisdiction over Defendants because Defendants are
`
`registered to, and do in fact, conduct substantial business throughout Illinois, including in this
`
`District, and Google maintains and uses two offices in Illinois.6 Moreover, Plaintiff’s claims arise
`
`out of Defendants’ conducting and transacting business in Illinois, and many of the actions giving
`
`rise to the claims asserted herein took place in this District.
`
`
`6
`See Our Offices, GOOGLE, https://about.google/intl/ALL_us/locations/?region=north-america&office=
`mountain-view (last visited Aug. 29, 2022).
`
`- 5 -
`
`

`

`Case 3:22-cv-02022 Document 1 Filed 08/30/22 Page 7 of 28 Page ID #7
`
`
`
`25.
`
`Venue is proper in this Court pursuant to 28 U.S.C. §1391 because many of the acts
`
`and transactions giving rise to this action occurred in this District and because Defendants:
`
`(a)
`
`are authorized to conduct business in this District and have intentionally
`
`availed themselves of the laws and markets within this District;
`
`(b)
`
`(c)
`
`conduct substantial business in this District; and
`
`are subject to personal jurisdiction in this District.
`
`FACTUAL BACKGROUND
`
`I.
`
`Google
`
`26.
`
`Google was launched in 1998 as a general online search engine. Founded by Larry
`
`Page and Sergey Brin, the corporation got its start by serving users web results in response to
`
`online queries. Google’s key innovation was its PageRank algorithm, which ranked the relevance
`
`of a webpage by assessing how many other webpages linked to it. In contrast with the technology
`
`used by rival search engines, PageRank enabled Google to improve the quality of its search results
`
`even as the web rapidly grew. While Google had entered a crowded field, by 2000, it had become
`
`the world’s largest search engine.
`
`27.
`
`Today, Google is ubiquitous across the digital world, serving as the infrastructure
`
`for core products and services online. It has grown and maintained its search engine dominance,
`
`such that “Googling” something is now synonymous with online search itself. The company is
`
`now also the largest provider of digital advertising, a leading web browser, a dominant mobile
`
`operating system, and a major provider of digital mapping, email, video hosting, cloud computing,
`
`and voice assistant services, alongside dozens of other offerings. Nine of Google’s products –
`
`Android, Chrome, Gmail, Google Search, Google Drive, Google Maps, Google Photos, Google
`
`Play Store, and YouTube – have more than a billion users each.
`
`- 6 -
`
`

`

`Case 3:22-cv-02022 Document 1 Filed 08/30/22 Page 8 of 28 Page ID #8
`
`
`
`28.
`
`Google established its position through acquisition, buying up successful
`
`technologies that other businesses had developed. In a span of 20 years, Google purchased well
`
`over 200 companies.
`
`29.
`
`Google is now one of the world’s largest corporations. For 2021, Google reported
`
`total consolidated revenues of over $257 billion – up 41% from 2020 – and more than $76 billion
`
`in net income.7
`
`II.
`
`YouTube
`
`30.
`
`YouTube was founded in 2005. It was the brainchild of Chad Hurley, Steve Chen,
`
`and Jawed Karim, who were all former employees of PayPal.
`
`31.
`
`According to YouTube’s founders, the idea was born at a dinner party in San
`
`Francisco, about a year earlier, in 2004. The trio was frustrated by how hard it was, at the time, to
`
`find and share video clips online. ‘“Video, we felt, really wasn’t being addressed on the Internet,’”
`
`said Chad Hurley in an early interview. ‘“People were collecting video clips on their cell phones
`
`. . . but there was no easy way to share [them].”’8
`
`32.
`
`By September 2005, YouTube had managed to get its first video with one million
`
`views. This was a Nike ad that went viral.
`
`33.
`
`34.
`
`Venture capitalists began pouring millions of dollars into YouTube in late 2005.
`
`In February 2006, YouTube for the first time added user profile personalization.
`
`The personalization of profiles feature was further refined in June 2006.
`
`
`7 Alphabet, Inc. Annual Report on Form 10-K for fiscal year ended Dec. 31, 2021 at 36,
`https://www.sec.gov/Archives/edgar/data/1652044/000165204422000019/goog-20211231.htm.
`
`8 Christopher McFadden, YouTube’s History and Its Impact on the Internet, INTERESTING ENGINEERING
`(May 20, 2021), https://interestingengineering.com/culture/youtubes-history-and-its-impact-on-the-
`internet.
`
`- 7 -
`
`

`

`Case 3:22-cv-02022 Document 1 Filed 08/30/22 Page 9 of 28 Page ID #9
`
`
`
`35.
`
`Just a few months later, in October 2006, Google acquired YouTube for $1.65
`
`billion. At the time, Google called YouTube, “the next step in the evolution of the Internet.”9
`
`36.
`
`37.
`
`In December 2009, YouTube’s automatic speech recognition service was launched.
`
`In July 2012, Google launched YouTube’s face-blurring tool, ostensibly to protect
`
`the anonymity of protesters and individuals engaging in civil disobedience around the world.10
`
`According to THE GUARDIAN,
`
`YouTube users who upload a video to the site are asked whether they want to apply
`a “Blur All Faces” option which will obscure all identities in the clip. Once faces
`have been obscured, YouTube creates two versions of the video, one without the
`blurring and one with. Users can decide whether to publish either or both of the
`videos. If they choose to delete the unblurred version, it will be removed
`permanently from Google’s servers.11
`
`III. Biometrics and Consumer Privacy
`
`38.
`
`“Biometrics” refers to technologies used to identify an individual based on unique
`
`physical characteristics. Common biometric identifiers include retina or iris scans, fingerprints,
`
`voiceprints, or hand or face geometry scans. One of the most prevalent uses of biometrics is facial
`
`recognition technology, which works by scanning an image for human faces, extracting facial
`
`feature data, and comparing them against information stored in a “faceprint database.” If a
`
`database match is found between the extracted facial data and the “biometric identifier” (i.e.,
`
`details about the face’s geometry), a person may be identified.
`
`
`9
`Id.
`
`10 Amanda Conway, Face blurring: when footage requires anonymity, YOUTUBE OFFICIAL BLOG (July
`18, 2012), https://blog.youtube/news-and-events/face-blurring-when-footage-requires/.
`
`11 Josh Halliday, Google introduces face-blurring to protect protesters on YouTube, THE GUARDIAN (July
`19,
`2012),
`https://www.theguardian.com/technology/2012/jul/19/face-blurring-technology-youtube-
`protestors#:~:text=Google%20introduces%20face%2Dblurring%20to%20protect%20protesters%20on%2
`0YouTube,This%20article%20is&text=Human%20rights%20activists%20and%20campaigners,blurring
`%20technology%20to%20the%20website.
`
`- 8 -
`
`

`

`Case 3:22-cv-02022 Document 1 Filed 08/30/22 Page 10 of 28 Page ID #10
`
`
`
`39.
`
`The recent sophistication of facial recognition software has generated many
`
`commercial applications of the technology, but has also raised serious privacy concerns about its
`
`massive scale, scope, and surreptitiousness.12 During a 2012 United States Senate hearing, Senator
`
`Al Franken noted that someone armed with a faceprint can find that person’s “name, . . . social
`
`networking account, . . . and can find and track [them] in the street, in the stores [they] visit, the
`
`Government buildings [they] enter, and the photos [their] friends post online.”13 Faceprints can
`
`be used even to identify protesters from afar and then “target them for selective jailing and
`
`prosecution[.]”14
`
`40.
`
`Unlike other identifiers such as Social Security or credit card numbers, which can
`
`be changed if compromised or stolen, biometric identifiers linked to a specific voice or face cannot.
`
`These unique, permanent, and immutable biometric identifiers, once exposed, leave victims with
`
`no means to prevent identity theft and unauthorized tracking. Recognizing this, the Federal Trade
`
`Commission urged companies using facial recognition technology to ask for consent from
`
`consumers before ever scanning and extracting biometric data from their digital photos.15
`
`Defendants have ignored this, failed to obtain user consent before launching their wide-spread
`
`facial recognition program, and violated millions of Illinois residents’ legal privacy rights.
`
`
`12 What Facial Recognition Technology Means for Privacy and Civil Liberties: Hearing Before the
`Subcomm. on Privacy Tech & the Law of the S. Comm. on the Judiciary, 112th Cong. 1 (July 18, 2012)
`(statement
`of
`Jennifer
`Lynch,
`Staff Attorney,
`Electronic
`Frontier
`Foundation),
`https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2134497.
`13 What Facial Recognition Technology Means for Privacy and Civil Liberties: Hearing Before the
`Subcomm. on Privacy Tech & the Law of the S. Comm. on the Judiciary, 112th Cong. 1 (July 18, 2012)
`(statement of Sen. Al Franken, Chairman, Subcomm. on Privacy, Tech. & the Law of the S. Comm. on the
`Judiciary). https://www.govinfo.gov/content/pkg/CHRG-112shrg86599/pdf/CHRG-112shrg86599.pdf.
`14 Id.
`15 See Facing Facts: Best Practices for Common Uses of Facial Recognition Technologies, Federal Trade
`Commission
`(Oct. 2012), http://www.ftc.gov/sites/default/files/documents/reports/facing-facts-best-
`practices-common-uses-facial-recognition-technologies/121022facialtechrpt.pdf.
`- 9 -
`
`

`

`Case 3:22-cv-02022 Document 1 Filed 08/30/22 Page 11 of 28 Page ID #11
`
`
`
`IV.
`
`Illinois’ Biometric Information Privacy Act
`
`41.
`
`BIPA was enacted in 2008. Under BIPA, a company may not “collect, capture,
`
`purchase, receive through trade, or otherwise obtain a person’s or a customer’s biometric
`
`identifier[16] . . . unless it first:
`
`(1) informs the subject . . . in writing that a biometric identifier . . . is being collected
`or stored;
`
`(2) informs the subject . . . in writing of the specific purpose and length of term for
`which a biometric identifier . . . is being collected, stored, and used; and
`
`(3) receives a written release executed by the subject of the biometric identifier . . . .”
`
`740 ILCS 14/15(b).
`
`42.
`
`BIPA also regulates how companies must handle Illinois consumers’ biometric
`
`data. See, e.g., 740 ILCS 14/15(c)-(d). For instance, BIPA prohibits selling, leasing, trading, or
`
`otherwise profiting from a person’s biometric data, 740 ILCS 14/15(c), and requires that
`
`companies develop a publicly available written policy establishing a retention schedule and
`
`guidelines for permanently destroying biometric data when the initial purpose for collecting such
`
`data has been satisfied or within three years of the individual’s last interaction with the company,
`
`whichever occurs first, 740 ILCS 14/15(a).
`
`V.
`
`Google Has a History of Violating BIPA
`
`43.
`
`In 2016, Google was sued for violating BIPA in connection with its Google
`
`Photos.17
`
`44.
`
`In general, plaintiffs in Rivera alleged that Google violated BIPA in the following
`
`manner:
`
`
`16 BIPA’s definition of “biometric identifier” expressly includes information collected about the geometry
`of the face (i.e., facial data obtained through facial recognition technology, like the data collected by
`Defendants about Plaintiff and the Class). See 740 ILCS 14/10.
`17 Class Action Complaint, Rivera v. Google LLC, No. 1:16-cv-02714 (N.D. Ill. Mar. 1, 2016), ECF No. 1.
`- 10 -
`
`

`

`Case 3:22-cv-02022 Document 1 Filed 08/30/22 Page 12 of 28 Page ID #12
`
`
`
`Google has created, collected and stored, in conjunction with its cloud-based
`“Google Photos” service, the “face templates” (or “face prints”) – highly detailed
`geometric maps of the face – of millions of users of the Google Photos service and
`hundreds of thousands of individuals who are not even enrolled in the Google
`Photos service. Google creates these templates using sophisticated facial
`recognition technology that extracts and analyzes data from the points and contours
`of faces that appear in photos taken on Google “Droid” devices and uploaded to the
`cloud-based Google Photos service. Each face template that Google extracts is
`unique to a particular individual, in the same way that a fingerprint or voiceprint
`uniquely identifies one and only one person.18
`
`45.
`
`46.
`
`Google settled Rivera by agreeing to pay $100 million.19
`
`Importantly, the Rivera settlement expressly covers only Google’s violation of
`
`BIPA with Google Photos.20 Nor could the Release in Rivera conceivably cover Defendants’
`
`ongoing BIPA violations with YouTube, under the long-standing “identical factual predicate”
`
`doctrine.21
`
`VI. Defendants Continue to Violate BIPA in Myriad Ways
`
`
`YouTube’s “Face Blur” Tool
`
`47.
`
`48.
`
`As noted, Defendants launched the “Face Blur” tool for YouTube in 2012.
`
`Unbeknownst to the average consumer, Defendants’ “Face Blur” tool relies on
`
`state-of-the-art facial recognition technology to scan videos, locate human faces, and create and
`
`store scans of face geometry.
`
`
`18 Second Amended Consolidated Complaint at 3, Rivera v. Google LLC, No. 1:16-cv-02714 (N.D. Ill.
`Mar. 7, 2017), ECF No. 63. Plaintiffs in Rivera also sued Google in Cook County, Illinois Circuit Court
`Chancery Division. See Class Action Complaint, Rivera v. Google LLC, No. 2019-CH-00990 (Ill. Ch. Ct.
`Jan. 24, 2019).
`
`19 See Jon Fingas, Google settles Photos facial recognition lawsuit for $100 million, ENGADGET (June 6,
`2022), https://www.engadget.com/google-photos-bipa-lawsuit-settlement-161237789.html.
`
`20 See Settlement Agreement at 9, Rivera v. Google LLC, No. 2019CH00990 (Ill. Ch. Ct. Apr. 13, 2022)
`(defining “Released Claims”).
`
`21 See Wal-Mart Stores, Inc. v. Visa U.S.A., Inc., 396 F.3d 96, 107 (2d Cir. 2005); Williams v. Gen. Elec.
`Cap. Auto Lease, Inc., 159 F.3d 266, 273 (7th Cir. 1998).
`
`- 11 -
`
`

`

`Case 3:22-cv-02022 Document 1 Filed 08/30/22 Page 13 of 28 Page ID #13
`
`
`
`49.
`
`Figure 1 below shows that the “Face Blur” tool is available as one of five tools on
`
`the YouTube “Video Editor” interface.
`
`Fig. 1
`
`
`
`50. When a YouTube video creator chooses to run the “Face Blur” tool on a video,
`
`Defendants scan the entire video to detect all unique faces within the video, as shown in Figure 2
`
`below.
`
`- 12 -
`
`
`
`

`

`Case 3:22-cv-02022 Document 1 Filed 08/30/22 Page 14 of 28 Page ID #14
`
`
`
`Fig. 2
`
`51.
`
`Once Defendants finish scanning the video, Defendants display all detected faces
`
`within the video and allow the creator to select which faces the creator would like to blur out in
`
`the video, as shown in Figure 3 below.
`
`Fig. 3
`
`
`
`52.
`
`After a creator selects a face from the list of all detected faces and applies the “Face
`
`Blur” tool, Defendants blur out the selected face throughout the duration of the YouTube video,
`
`as shown in Figure 4 below.
`
`- 13 -
`
`

`

`Case 3:22-cv-02022 Document 1 Filed 08/30/22 Page 15 of 28 Page ID #15
`
`
`
`Fig. 4.
`
`
`
`53.
`
`The underlying computer code depicted in Figure 5 below demonstrates that when
`
`a YouTube user employs Defendants’ “Face Blur” tool, Defendants capture and store on their
`
`servers scans of face geometry from all detected faces, or biometric information and identifiers.
`
`This is seen by the unique “faceId” and associated image representation of the faces in the code
`
`pictured in Figure 5 below. Defendants use the “faceId” to match with the specific facial geometry
`
`captured and stored on Defendants’ servers. Then, when users select a “faceId,” Defendants blur
`
`any matching facial geometry stored on Defendants’ servers.
`
`- 14 -
`
`

`

`Case 3:22-cv-02022 Document 1 Filed 08/30/22 Page 16 of 28 Page ID #16
`
`
`
`Fig. 5.
`
`
`
`54.
`
`It appears that Defendants are only storing the detected faces for a total of four
`
`hours because each detected face sent to Defendants’ servers comes with an expiry date within a
`
`few hours of running the face detection. This is shown in Figure 6 below:
`
`Fig. 6.
`
`
`
`55.
`
`However, Defendants are actually storing the scan of face geometry for a longer
`
`period of time, and possibly permanently.
`
`56.
`
`This is demonstrated by the fact that, when the “Face Blur” tool is run multiple
`
`times on the same video, the previously stored result is provided to the user without actually
`
`- 15 -
`
`

`

`Case 3:22-cv-02022 Document 1 Filed 08/30/22 Page 17 of 28 Page ID #17
`
`
`
`rerunning the tool again. This is true even when the “Face Blur” tool is run multiple weeks after
`
`initially running the “Face Blur” tool – even though Defendants represent to no longer have access
`
`to the detected faces after a few hours. See Figure 7 below.
`
`
`
`Fig. 7.
`
`
`
`57.
`
`Consequently, Defendants permanently store scans of face geometry or biometric
`
`information so that YouTube users do not need to re-run the “Face Blur” tool.
`
`58.
`
`Defendants do not disclose this collection and storage of biometric information or
`
`identifiers anywhere, neither in their Terms of Service or elsewhere.
`
`59.
`
`Even if Defendants did disclose the collection and storage of biometric information
`
`or identifiers to the creator of the video who used the “Face Blur” tool, it would still be a legally
`
`insufficient disclosure to the individual whose image was captured and stored within the video.
`
`60.
`
`Accordingly, Defendants have violated and continue to violate the rights of Plaintiff
`
`and the Class under BIPA.
`
`- 16 -
`
`

`

`Case 3:22-cv-02022 Document 1 Filed 08/30/22 Page 18 of 28 Page ID #18
`
`
`
`YouTube’s Thumbnail Generator
`
`61.
`
`Defendants also include with YouTube a feature that at first auto-generates
`
`photographic thumbnails (screenshots from an uploaded video) and allows creators to choose their
`
`own thumbnails for their videos.22
`
`62.
`
`It is common knowledge that thumbnails with faces, especially faces with more
`
`expression, generate more clicks and views. As such, Defendants have a huge financial incentive
`
`to auto-generate thumbnails that contain faces – especially faces that contain more expression.
`
`63.
`
`Upon information and belief, in order to generate optimal thumbnails, Defendants
`
`scan all videos uploaded to YouTube for faces at the time the videos are uploaded, and then use
`
`this face data to auto-generate thumbnails that contain faces, and especially faces with more
`
`expression.
`
`64.
`
`An experiment conducted by Plaintiff’s counsel verifies Plaintiff’s allegations.
`
`Three videos of approximately 60 seconds long were uploaded to YouTube. Each video contained
`
`a face within the video for less than two seconds. The remainder of the video contained other
`
`content (e.g., cars driving, scenery out of the window, a house tour). In each of the three videos,
`
`Defendants’ thumbnail generator, without prompting, auto-generated and applied as the main
`
`thumbnail a face that was only in the video for less than two seconds. A sample illustrating this is
`
`in Figure 8 below:
`
`
`22 Add video
`thumbnails on YouTube, YOUTUBE, https://support.google.com/youtube/answer/
`72431?hl=en#zippy=%2Cwhy-are-my-custom-thumbnails-turned-off (last visited Aug. 29, 2022).
`
`- 17 -
`
`

`

`Case 3:22-cv-02022 Document 1 Filed 08/30/22 Page 19 of 28 Page ID #19
`
`
`
`Fig. 8.
`
`
`
`65.
`
`This indicates that, as part of the Thumbnail Generator, Defendants have a software
`
`program that has been trained to detect faces. This software program would require scans of facial
`
`geometry from YouTube videos as a necessary input in order to train this software. Further, it is
`
`likely that the software is continually being optimized by gathering more facial geometry data from
`
`YouTube videos.
`
`66.
`
`Defendants’ own research from 2021 also indicates that it has analyzed millions of
`
`videos to detect faces. The research states that it used software “similar to the Google Cloud Face
`
`Detection API” to scan and detect faces, follow faces over the course of the video, and label facial
`
`expressions in the videos.23
`
`67.
`
`Based on the above, it is reasonable and plausible to assume that Defendants are
`
`scanning each uploaded YouTube video for faces, and auto-generating thumbnails with expressive
`
`
`23 Understanding Contextual Facial Expressions Across the Globe, GOOGLE AI BLOG (May 24, 2021),
`https://ai.googleblog.com/2021/05/understanding-contextual-facial.html
`
`- 18 -
`
`

`

`Case 3:22-cv-02022 Document 1 Filed 08/30/22 Page 20 of 28 Page ID #20
`
`
`
`facial expressions. During this process, Defendants are scanning, detecting, and collecting facial
`
`geometry within each YouTube video, and storing the metadata associated with the video on
`
`Defendants’ servers.
`
`Defendants Never Require Users to Acknowledge Their Biometric Data Collection
`Practices, Never Obtains Their Express Written Consent to Collect the Same, and,
`Instead, Hide the Fact that They Systematically Collect Users’ Biometrics
`
`68.
`
`Since Defendants’ “Face Blur” tool debuted in 2012, and for as long as Defendants’
`
`thumbnail creator has been available, Defendants have never disclosed their collection and storage
`
`of Plaintiff’s and the Class’ biometric data.
`
`69.
`
` First, Defendants do not directly or indirectly inform users that they collect,
`
`capture, and store faceprints from users. Nor do Defendants require users to acknowledge their
`
`collection and storage of their biometric data, much less obtain a written release from them before
`
`collecting their faceprints. Instead, Defendants market their “Face Blur” tool as protecting
`
`“anonymity” and their thumbnail creator as simply a beneficial feature for YouTube video
`
`creators. Defendants never give any indication that use of these features would come at the cost
`
`of users’ biometric privacy rights.
`
`70.
`
`Secondly, and compounding these problems and their violations of BIPA,
`
`YouTube’s website does not have a written, publicly available policy identifying its biometrics
`
`retention schedule, nor guidelines for permanently destroying users’ biometric identifiers when
`
`they are no longer needed.
`
`71.
`
`By and through these actions, Defendants not only disregarded their users’ privacy
`
`rights, but they also violated their statutorily protected rights to control the collection, use, and
`
`storage of their sensitive biometric data.
`
`- 19 -
`
`

`

`Case 3:22-cv-02022 Document 1 Filed 08/30/22 Page 21 of 28 Page ID #2