Campbell et al v. Facebook Inc.

Filing 162

Exhibits in Support of 147 Administrative Motion to File Under Seal and Documents in Support of Facebooks Opposition to Plaintiffs Motion for Class Certification (Dkt. 149) filed by Facebook Inc.. (Attachments: # 1 Replacement for Dkt. 147-1 (Declaration of Nikki Stitt Sokol In Support Of Defendant Facebook, Inc.s Administrative Motion to File Documents in Support of its Opposition to Plaintiffs Motion for Class Certification Under Seal), # 2 Replacement for Dkt. 147-2 ([Proposed] Order Authorizing the Filing of Documents Under Seal), # 3 Replacement for Dkt. 147-5 ((Exhibit 3) Unredacted Chorba Declaration Motion to Seal), # 4 Replacement for Dkt. 147-6 ((Exhibit 4) Redacted Chorba Declaration Motion to Seal), # 5 Replacement for Dkt. 149-1 (Redacted Chorba Declaration Opposition to Class Certification), # 6 Replacement for Dkt. 149-7 (Redacted Expert Report of Dr. Catherine Tucker Opposition to Class Certification), # 7 Replacement for Dkt. 154-11 ((Exhibit 66) Unredacted Tucker Expert Report Motion to Seal), # 8 Replacement for Dkt. 154-12 ((Exhibit 67) Redacted Tucker Expert Report Motion to Seal), # 9 Replacement for Dkt. 155-1 (Unredacted Appendix of Evidence (Part 1 of 19) Motion to Seal), # 10 Replacement for Dkt. 156-9 (Unredacted Appendix of Evidence (Part 19 of 19) Motion to Seal), # 11 Replacement for Dkt. 157-1 (Redacted Appendix of Evidence (Part 1 of 13) Opposition to Class Certification), # 12 Replacement for Dkt. 157-13 (Unredacted Appendix of Evidence (Part 13 of 13)) Motion to Seal))(Chorba, Christopher) (Filed on 1/22/2016) Modified on 1/22/2016 (vlkS, COURT STAFF).

Download PDF
Replacement for Dkt. 157-1 1 2 3 4 5 6 7 8 9 10 11 12 13 GIBSON, DUNN & CRUTCHER LLP JOSHUA A. JESSEN, SBN 222831 JJessen@gibsondunn.com JEANA BISNAR MAUTE, SBN 290573 JBisnarMaute@gibsondunn.com PRIYANKA RAJAGOPALAN, SBN 278504 PRajagopalan@gibsondunn.com ASHLEY M. ROGERS, SBN 286252 ARogers@gibsondunn.com 1881 Page Mill Road Palo Alto, California 94304 Telephone: (650) 849-5300 Facsimile: (650) 849-5333 GIBSON, DUNN & CRUTCHER LLP CHRISTOPHER CHORBA, SBN 216692 CChorba@gibsondunn.com 333 South Grand Avenue Los Angeles, California 90071 Telephone: (213) 229-7000 Facsimile: (213) 229-7520 Attorneys for Defendant FACEBOOK, INC. 14 UNITED STATES DISTRICT COURT 15 NORTHERN DISTRICT OF CALIFORNIA 16 17 18 OAKLAND DIVISION MATTHEW CAMPBELL and MICHAEL HURLEY, 19 Plaintiffs, 20 v. 21 FACEBOOK, INC., 22 23 24 Defendant. Case No. C 13-05996 PJH PUTATIVE CLASS ACTION APPENDIX OF EVIDENCE IN SUPPORT OF DEFENDANT FACEBOOK, INC.’S OPPOSITION TO PLAINTIFFS’ MOTION FOR CLASS CERTIFICATION HEARING: Date: March 16, 2016 Time: 9:00 a.m. Place: Courtroom 3, 3rd Floor The Honorable Phyllis J. Hamilton 25 26 27 28 Gibson, Dunn & Crutcher LLP APPENDIX OF EVIDENCE IN SUPPORT OF DEFENDANT FACEBOOK, INC.’S OPPOSITION TO PLAINTIFFS’ MOTION FOR CLASS CERTIFICATION Case No. C 13-05996 PJH 1 2 3 4 5 6 7 8 9 10 11 12 GIBSON, DUNN & CRUTCHER LLP JOSHUA A. JESSEN, SBN 222831 JJessen@gibsondunn.com JEANA BISNAR MAUTE, SBN 290573 JBisnarMaute@gibsondunn.com PRIYANKA RAJAGOPALAN, SBN 278504 PRajagopalan@gibsondunn.com ASHLEY M. ROGERS, SBN 286252 ARogers@gibsondunn.com 1881 Page Mill Road Palo Alto, California 94304 Telephone: (650) 849-5300 Facsimile: (650) 849-5333 GIBSON, DUNN & CRUTCHER LLP CHRISTOPHER CHORBA, SBN 216692 CChorba@gibsondunn.com 333 South Grand Avenue Los Angeles, California 90071 Telephone: (213) 229-7000 Facsimile: (213) 229-7520 Attorneys for Defendant FACEBOOK, INC. 13 14 UNITED STATES DISTRICT COURT 15 NORTHERN DISTRICT OF CALIFORNIA 16 OAKLAND DIVISON 17 MATTHEW CAMPBELL and MICHAEL HURLEY, 18 Plaintiffs, 19 v. 20 FACEBOOK, INC., 21 Defendant. 22 23 24 25 26 27 28 Gibson, Dunn & Crutcher LLP APP. 1 Case No. C 13-05996 PJH DECLARATION OF CHRISTOPHER CHORBA IN SUPPORT OF DEFENDANT FACEBOOK, INC.’S OPPOSITION TO PLAINTIFFS’ MOTION FOR CLASS CERTIFICATION I, Christopher Chorba, declare as follows: 1 1. 2 I am an attorney admitted to practice law before this Court. I am a partner in the law 3 firm of Gibson, Dunn & Crutcher LLP, and I am one of the attorneys responsible for representing 4 Defendant Facebook, Inc. (“Facebook”) in the above-captioned action. I submit this declaration in 5 support of Facebook’s Opposition to Plaintiffs’ Motion for Class Certification (Dkt. 138). Unless 6 otherwise stated, the following facts are within my personal knowledge and, if called and sworn as a 7 witness, I could and would testify competently to these facts. 8 I. 9 10 Demonstratives 2. and challenged practices. * a. Attached as Exhibit A is a chart summarizing a number of individualized issues 11 12 concerning the named Plaintiffs and some putative class members. b. Attached as Exhibit B is a graphical representation of the steps required to send 13 14 and receive a Facebook message with a URL preview attachment. 15 c. Attached as Exhibit C are graphical representations of the individualized inquiries 16 related to ascertainability. 17 d. Attached as Exhibit D are charts summarizing the variability for the challenged 18 practices. 19 20 21 22 23 24 Attached as Exhibits A–D are demonstrative graphics regarding the named plaintiffs 3. Facebook and its messaging service have often been the subject of public news reports, blog posts, and other publications. Attached as Exhibit E is a chart summarizing seventyseven publicly available online publications, including, inter alia, news reports, articles, editorials, and Facebook developer documentation, published between May 6, 2009 and August 7, 2013. Attached as Exhibits F, G, H, I, J, and K are the corresponding seventy-seven publications, arranged by Bates numbers FB000000066 to FB000000424 and produced by Facebook during this litigation. 25 26 27 * For the Court’s convenience, and to avoid duplication in the numbering of the exhibits submitted by Plaintiffs, Facebook has used letters rather than numbers to designate its exhibits. 28 Gibson, Dunn & Crutcher LLP APP. 2 1 II. Discovery Requests And Responses From Plaintiffs A. 2 3 4 4. 5. Attached as Exhibit M is a true and correct copy of relevant excerpts of the deposition transcript of Plaintiff Michael Hurley on July 9, 2015. 7 8 Attached as Exhibit L is a true and correct copy of relevant excerpts of the deposition transcript of Plaintiff Matthew Campbell on May 19, 2015. 5 6 Plaintiffs’ Deposition Testimony 6. Attached as Exhibit N is a true and correct copy of relevant excerpts of the deposition transcript of Mr. David Shadpour on October 1, 2015. B. 9 10 7. Plaintiffs’ Written Discovery Responses Attached as Exhibit O is a true and correct copy of Plaintiff Campbell’s Corrected 11 Objections and Responses to Defendant Facebook, Inc.’s First Set of Interrogatories, dated April 2, 12 2015. As these responses reflect, Mr. Campbell has sent or received at least 13 containing URLs between the time he filed this action (December 30, 2013), and the date of his 14 responses (April 2, 2015). 15 8. Facebook messages Attached as Exhibit P is a true and correct copy of Plaintiff Hurley’s Objections and 16 Responses to Defendant Facebook, Inc.’s First Set of Interrogatories, dated April 1, 2015. As these 17 responses reflect, Mr. Hurley has sent or received at least 18 between the time he filed this action (December 30, 2013), and the date of his responses (April 1, 19 2015). 20 9. Facebook messages containing URLs Attached as Exhibit Q is a true and correct copy of (Former) Plaintiff Shadpour’s 21 Corrected Objections and Responses to Defendant Facebook, Inc.’s First Set of Interrogatories, dated 22 April 2, 2015. As these responses reflect, Mr. Shadpour has sent or received at least 23 messages containing URLs between the time he filed this action (January 21, 2014), and the date of 24 his responses (April 2, 2015). 25 10. Facebook On April 10, 2015, Plaintiffs supplemented their responses to Facebook’s 26 Interrogatories through a letter from counsel (David Rudolph). In particular, Plaintiffs supplemented 27 their responses to Facebook’s Interrogatory No. 5 to describe the manner in which they learned of the 28 Gibson, Dunn & Crutcher LLP APP. 3 1 2 (Id.) At the time of the Responses, Facebook determined that 3 4 5 6 (Id.) Facebook produced documents related to its responses regarding the 19 messages. (Id. at 18 & Ex. A) 17. Facebook also analyzed these messages to determine which of the messages (if any) 7 had a possibility of incrementing a social plugin count on a third-party website. Although Facebook 8 does not possess records to determine whether a particular third-party webpage displayed a social 9 plugin count at the time Plaintiffs’ selected messages were either sent or received, the Internet 10 Wayback Machine (https://archive.org/web/) is a “reliable” resource that Plaintiffs’ technical expert, 11 Dr. Jennifer Golbeck, uses “pretty frequently” to view archived webpages. (Ex. EE, Golbeck Depo. 12 Tr., at 20:7-21:3.) 13 18. For each of the remaining twelve messages selected by Plaintiffs and for which a share 14 object was created, the Internet Wayback Machine revealed that for the 10 of 12 messages that did 15 have a share object, there was no corresponding social plugin on the websites referenced by the URLs 16 in Plaintiffs’ messages at or near the time the messages were sent. For example, on 17 18 19 20 21 22 23 Thus, 10 of the 19 messages identified by Plaintiffs had a share object but did not have a corresponding social plugin on the third-party website. 19. For 1 of the 12 messages that did have a share object, the Internet Wayback Machine did not have the webpage archived. That message was sent by 24 25 26 20. 27 Gardner. The remaining message was sent from Plaintiff Hurley to Plaintiffs’ counsel Melissa 28 Gibson, Dunn & Crutcher LLP APP. 6 1 III. Other Discovery Issues A. 2 3 21. 4 6 7 8 9 (Dkt. 138 at 20-21.) 22. 11 12 13 14 15 18 19 20 21 In fact, Facebook’s counsel never told Plaintiffs’ counsel that Facebook did not have a “dedicated team of privacy professionals.” On the contrary, Facebook specifically denied Plaintiffs’ request to admit that there was no such team, and indeed there is such a team. Attached as Exhibit X is a true and correct copy of Defendant Facebook, Inc.’s Responses and Objections to Plaintiffs’ First Set of Requests for Admission dated June 29, 2015. 23. 16 17 In their Motion, Plaintiffs assert as follows: Discovery also demonstrates that Facebook’s public-facing statements about “procedural safeguards” for ensuring user privacy in product development are false. Facebook has represented, inter alia, in its filings with the Security and Exchange Commission that it has “a dedicated team of privacy professionals who are involved in new product and feature development from design through launch” and who conduct “ongoing review and monitoring of the way data is handled by existing features and applications.” However, when asked to produce documents sufficient to identify the individuals comprising this “dedicated team,” Facebook responded that none existed. 5 10 Facebook’s “Public-Facing Statements” and “Dedicated Team of Privacy Professionals” Rather, Facebook’s counsel simply confirmed that, in response to a document request, there was not a “specific list.” Plaintiffs’ request sought “documents” regarding “the ‘dedicated team of privacy professionals’ identified on page 8 of Your Form 10-K for fiscal year ending December 31, 2013.” (Dkt. 138-4, Ex. 31.) Facebook responded by explaining that it did not have a document responsive to Request No. 29, listing members of its internal privacy team. Plaintiffs even misstated the correspondence among counsel by omitting the bolded portion below in their brief: With respect to Request No. 29, please be advised that there is no specific list of the ‘dedicated team of privacy professionals’ referenced in the Request, but we have already agreed to conduct a reasonable search for non-privileged documents sufficient to identify Facebook’s current and former employees who may possess knowledge relevant to the practice challenged in this action, and we also have identified witnesses with relevant knowledge in Facebook’s Initial Disclosures and responses to Plaintiffs’ Interrogatories. 22 23 24 25 26 Plaintiffs attached Facebook’s complete response to the request as Exhibit 32 (Dkt. 138-4, Ex. 27 32). 28 Gibson, Dunn & Crutcher LLP APP. 7 1 2 B. 24. Plaintiffs’ Expanded Proposed Class Definition Exceeds The “Relevant Time Period” For Discovery Plaintiffs’ Consolidated Amended Complaint identified the following proposed class: 3 “All natural-person Facebook users located within the United States who have sent or received 4 private messages that included URLs in their content, from within two years before the filing of this 5 action up through and including the date when Facebook ceased its practice,” which Plaintiffs alleged 6 to be “at some point after it was exposed in October 2012.” (Dkt. 25 ¶ 59 & n.3.) 7 25. In their Motion for Class Certification, Plaintiffs now seek to certify a proposed class 8 of all “Facebook users located within the United States who have sent, or received from a Facebook 9 user, private messages that included URLs in their content (and from which Facebook generated a 10 URL attachment), within two years before the filing of this action up through the date of 11 certification of the class.” (Dkt. 138 at 10 (emphasis added).) In other words, Plaintiffs have now 12 expanded their proposed class by over three years. 13 26. Plaintiffs’ new proposed class definition extends well beyond the relevant time period 14 to which the parties expressly agreed for discovery. On April 7, 2015, Hank Bates, counsel for 15 Plaintiffs, proposed that the “Relevant Time Period” for “producing documents” should be April 1, 16 2010, to the date of filing the action, December 30, 2013. Attached as Exhibit Y is a true and correct 17 copy of Mr. Bates’ letter dated April 7, 2015. 18 27. After some further discussions between the parties, Facebook agreed to this time 19 period in letters dated May 13 and June 12, 2015. Attached as Exhibits Z and AA are true and 20 correct copies of these letters. 21 28. Regarding the production of source code, the parties agreed (and stipulated, see 22 Dkt. 90) to a slightly different time period—September 1, 2009 to December 31, 2012—reflecting the 23 fact that Plaintiffs had alleged that the challenged practice had ceased “at some point after it was 24 exposed in October 2012.” (Dkt. 25 ¶ 59 & n.3.) 25 29. Additionally, during depositions of Facebook’s witnesses, counsel for Plaintiffs 26 repeatedly limited questions to the time period of “2010 to 2012” or “2010 to 2013.” Attached as 27 Exhibits BB and CC are true and correct copies of excerpts of the deposition transcripts of Facebook 28 Gibson, Dunn & Crutcher LLP APP. 8 1 witnesses, Jiakai Liu and Ray He, dated June 30, 2015 and September 25, 2015, respectively, 2 reflecting, inter alia, a handful of those questions. 3 C. 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 30. Fernando Torres’ Expert Report And The Information He Claims That He Needs To Complete His Damages Analysis Plaintiffs’ proposed damages expert, Mr. Fernando Torres, testified that, in order to complete his damages analysis, he needed additional information that is distinct from Plaintiffs’ previous damages discovery requests— which they represented were “critical to establishing” their damages theory. Attached as Exhibit DD is a true and correct copy of relevant excerpts of the deposition transcript of Mr. Fernando Torres on December 18, 2015. 31. In support of prior discovery motions, Plaintiffs argued that they would be “unduly prejudice[d]” without “discovery relevant to damages in this action.” (Dkt. 112 at 2; see also Dkt. 109 at 2, 4 (arguing that “[w]ithout discovery into the revenue Facebook has generated . . . Plaintiffs will be hampered in formulating a class-wide damages theory”).) Plaintiffs represented that the discovery they sought was “critical to establishing” their damages theory and that “expert analysis of the [] information sought” would allow them to “accurately model the profits attributable to the challenged conduct.” (Dkt. 112 at 2-3.) And they also argued that the damages discovery sought was “directly relevant to the issues of damages suffered by the class as well as the appropriate injunctive relief . . . and [was] . . . necessary for Plaintiffs to fashion a theory of class-wide relief for their class certification briefing.” (Dkt. 109 at 2, 4.) 32. In light of these and other arguments, Plaintiffs received a 30-day extension of the briefing schedule (Dkt. 117) and successfully compelled Facebook to produce extremely broad discovery (Dkt. 130, 136.). 33. In his expert report, however, Mr. Torres cited only 7 of the thousands of documents produced by Facebook during the course of this litigation. (Dkt. 138-4, Ex. 33.) He also asserted in his report that he needed other information from Facebook: “with additional information, including production from Facebook, and inputs, these conclusions [in the Report] could be refined.” (Dkt. 138-4, Ex. 33, ¶ 11 n.12.) In the final paragraph of his report, Mr. Torres explained, “With quantitative data on the number of affected ‘Like’ counts, and identification of the affected URLs, it Gibson, Dunn & Crutcher LLP APP. 9 1 prepare their motion for class certification” was “prejudice[ed]” by Facebook’s alleged “delay[s] 2 providing relevant discovery in this matter.” (Dkt. 138-3, ¶ 2.) More specifically, he claims that 3 Facebook “delayed production of its source code by over five months . . . and [] failed to produce a 4 significant number of documents responsive to Plaintiffs’ document requests” in a timely manner. 5 (Id.) 6 38. Mr. Rudolph does not explain that this Court already was presented with these 7 arguments on two separate occasions. After considering Facebook’s Opposition to Plaintiffs’ Motion 8 to Enlarge Time and Extend Deadlines (Dkt. 114) and the supporting Declaration of Joshua Jessen 9 (Dkt. 114-1), which rebutted similar assertions from Plaintiffs’ counsel, this Court ruled that the “90- 10 day extension sought by plaintiffs would unnecessarily delay the case,” and instead ordered a 30-day 11 extension. (Dkt. 117; see also Dkt. 113-1 at 13.) 12 39. Several weeks later, Plaintiffs filed a Renewed Motion to Continue, attempting to 13 revisit the issue and arguing that Facebook “delayed [] providing relevant discovery, including by 14 failing to produce a significant proportion of relevant and responsive documents until October 13, 15 and October 28.” (Dkt. 134-1.) Once again, Facebook responded to Plaintiffs’ false assertions and 16 corrected the record. (Dkt. 135, 135-1.) This Court denied Plaintiffs’ motion. (Dkt. 136.) 17 40. Mr. Rudolph’s most recent declaration (Dkt. 138-3) again argues that Facebook 18 “delayed” production of its source code, “delayed” producing a significant portion of documents until 19 October 13-28, 2015, and “delayed” producing additional documents until November 3-7, 2015. 20 (Dkt. 138-3, ¶¶ 2–5.) Facebook already refuted the first two assertions were before the Court. (See 21 Dkt. 114-1 ¶¶ 8–36; 135-1 ¶¶ 2–10.) On Mr. Rudolph’s last point, he fails to mention that 22 Facebook’s November productions were in response to Plaintiffs’ Motion to Compel (Dkt. 112), 23 Magistrate Judge James’ Order on October 14, 2015 (Dkt. 130), and this Court’s Order on 24 November 3, 2015. (Dkt. 136.) In other words, the productions were the result of Plaintiffs’ motions 25 to compel. Facebook produced all responsive documents it could locate after a reasonable search in a 26 timely manner. Although Mr. Rudolph is correct to point out that the November 7 productions were 27 significant in volume, this was through no fault of Facebook—it had repeatedly warned Plaintiffs that 28 Gibson, Dunn & Crutcher LLP APP. 12 1 their requests were extremely overbroad and would yield many irrelevant documents, and Facebook 2 undertook extensive efforts to try to reach a reasonable compromise. (Dkt. 131-1.) For example, 3 Facebook offered to provide Plaintiffs with representative documents for certain of Plaintiffs’ 4 requests, but Plaintiffs rejected all offers for compromise and continued to litigate these issues. 5 (Dkt. 131-1, Ex. 1.) 6 41. Contrary to Mr. Rudolph’s declaration, Facebook’s production was substantially 7 complete as of September 30, 2015, with respect to the documents Facebook had agreed to produce at 8 that point. Productions after this date were primarily in response to Plaintiffs’ Motion to Compel 9 (Dkt. 112, 113, 122), which were not even decided until after September 30. (See Dkt. 130, 136.) 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 IV. Authentication Of Remaining Exhibits 42. Attached as Exhibit EE is a true and correct copy of excerpts of the deposition transcript of Dr. Jennifer Golbeck (dated December 16, 2015). 43. transcript of 44. transcript of 45. transcript of 46. Attached as Exhibit FF is a true and correct copy of excerpts of the deposition (dated August 7, 2015). Attached as Exhibit GG is a true and correct copy of excerpts of the deposition (dated August 10, 2015). Attached as Exhibit HH is a true and correct copy of excerpts of the deposition (dated August 11, 2015). Attached as Exhibit II is a true and correct copy of excerpts of the deposition transcript of Ray He (dated October 28, 2015). 47. Attached as Exhibit JJ is a true and correct copy of excerpts of the deposition transcript of Michael Adkins (dated October 28, 2015). 48. Attached as Exhibit KK is a true and correct copy of a document that begins with Bates number FB000006429, which Facebook produced during this litigation. 25 26 27 28 Gibson, Dunn & Crutcher LLP APP. 13 1 49. Attached as Exhibits LL are true and correct copies of certain Google Analytics data 2 that begins with Bates numbers FB000009906 and FB000009914, respectively, and which Facebook 3 produced to Plaintiffs during discovery. 4 I declare under penalty of perjury under the laws of the United States of America and the 5 State of California that the foregoing is true and correct, and that I executed this Declaration in Los 6 Angeles, California, on January 15, 2016. 7 8 /s/ Christopher Chorba Christopher Chorba 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 Gibson, Dunn & Crutcher LLP APP. 14 EXHIBIT A APP. 15 App. 16-17 Filed Under Seal EXHIBIT B APP. 18 App. 19 Filed Under Seal EXHIBIT C APP. 20 App. 21-29 Filed Under Seal EXHIBIT D APP. 30 App. 31-37 Filed Under Seal EXHIBIT E APP. 38 Exhibit E: Evidence of Implied Consent Source Bates Number Publication Summary 1 FB000000079 Josh Constine, Facebook Allows Users to Upgrade to the New Messages Product, Why You Should, Inside Facebook (January 10, 2011), available at http://www.insidefacebook.com/2011/01/1 0/allows-upgrade-new-messages/. This article describes how a Facebook Messages product update “automatically filters non-essential communications into an Other folder, allowing the main inbox to show only important messages [and] also routes sent messages to whichever device or interface Facebook deems is the most convenient for the recipient, whether that’s Chat, Messages, SMS, or email.” The author praises the storage capacity and searchability of Facebook’s Messages product, but he writes that some “important Messages may be being filtered out.” The author concludes that “[o]verall, Messages will help most Facebook users [because] [i]t anticipates the shift to using multiple devices and interfaces to conduct a single conversation . . . declutters the inbox by removing spammy and low-value Page and Event updates . . . [and] improves one of the core uses of Facebook—instant communication with friends.” 2 FB000000156 Jiakai Liu, Inside Facebook Messages’ Application Server, Facebook (April 28, 2011), available at https://www facebook.com/notes/facebookengineering/inside-facebook-messagesapplication-server/10150162742108920/. In this Facebook Note, Facebook software engineer Jiakai Liu describes the technical details of the infrastructure behind Facebook’s Messages product, including the internals of the application server used to manage this infrastructure. He writes that the server will run “pre- and post-processing as needed, and determin[e] the folder and thread where the message should be routed based on a number of signals.” He also indicates that “[w]hen reading messages, the server gets various statistics about the user’s mailbox, like its capacity; number of messages, threads and replies; and the number of friends with whom the user has interacted. It also gets folder statistics and attributes, the list of threads by various search criteria (folder, attributes, authors, keywords, and so forth), and thread attributes and the individual messages in the thread).” He also discloses that Facebook’s Messages product supports “full text search” by maintaining a “reverse index from keywords to matched messages.” He describes how Facebook “parse[s]” and “convert[s]” a new messages when it arrives, and writes that “[a]ll messages, including chat history, email, and SMS, are indexed in real time.” He concludes by noting that Facebook did testing of the new application server software via a “dark launch,” but that Facebook would “continue to roll out the new Messages system to all our users.” 3 FB000000163 This Internet Archive capture is Facebook’s developer documentation describing how to implement the “Like” button Facebook, Facebook Developer Documentation: “Like” Button, Facebook - and “FAQs” about that implementation. It discloses that the number shown on a “Like” button “is the sum of [items Internet Archive Wayback Machine (March including] [t]he number of inbox messages containing this URL as an attachment.” 7, 2011), available at http://web.archive.org/web/201103072139 24/http://developers facebook.com/docs/ref erence/plugins/like/. 4 FB000000166 This Internet Archive capture is Facebook’s developer documentation describing how to implement the “Like” button Facebook, Facebook Developer Documentation: “Like” Button, Facebook- and “FAQs” about that implementation. It discloses that the number shown on a “Like” button “is the sum of [items Internet Archive Wayback Machine including] [t]he number of inbox messages containing this URL as an attachment.” (October 19, 2012), available at http://web.archive.org/web/201210190956 52/http://developers facebook.com/docs/ref erence/plugins/like/. Page 1 of 22 APP. 39 Exhibit E: Evidence of Implied Consent Source Bates Number Publication Summary 5 FB000000173 Elliot Schrage, Proposed Updates to our Governing Documents, Facebook (November 21, 2012), available at http://newsroom.fb.com/news/2012/11/pro posed-updates-to-our-governingdocuments/. In this News Post, a Facebook employee writes that Facebook is “proposing some updates to two documents which govern our site: our Data Use Policy, which explains how we collect and use data when people use Facebook, and our Statement of Rights and Responsibilities (SRR), which explains the terms governing the use of our services.” He notes that Facebook would continue to post “significant” changes to the documents for review and comment and provide “additional notification mechanisms” informating users of any changes to them. He also summarizes certain updates to the Data Use Policy, including 1) “[n]ew tools for managing your Facebook Messages - replacing the ‘Who can send you Facebook messages’ setting with new filters for managing incoming messages”; 2) changes to how Facebook refers to certain products; 3) reminders about what’s visible to others on Facebook; and 4) tips for managing timelines. He concludes by encouraging readers to read the proposed changes to the documents and leave comments. 10 FB000000240 Here, the author asks, “[D]id you know that the number of Likes is made up of more than just people clicking Like?” Thomas McMahon, Facebook’s Like Number is More than Just People Clicking Quoting Facebook’s developer site, he writes that “[t]he number shown on a Like button is the sum of: ‘[t]he number Like, Twister MC (September 8, 2011), of likes for the URL; [t]he number of shares for the URL – This includes copy/pasting a link back to Facebook; [t]he available at number of likes and comments on stories on Facebook about the URL; [and] [t]he number of inbox messages http://www.twistermc.com/36579/facebook- containing the URL as an attachment.’” Providing an example, the author notes that if you have a blog post with a like/. “Like” button social plug in, the “Like number showing would include: . . . The number of times someone has sent the blog post URL to a friend via Facebook’s messaging system.'” He concludes that “[b]asically, anytime that blog post URL is active on Facebook, a Like is added to the count.” 6 FB000000204 Nicholas Carlson, The Truth About The Latest Facebook Privacy Scare Everyone Is Talking About, Business Insider (October 4, 2012), available at http://www.businessinsider.com/facebookprivate-messages-likes-2012-10/. After reporting that “[w]hen a Facebook user sends a link to a Web page via a private Facebook message, that Web page will get an extra ‘Like,’ if it is a Facebook-‘Like’able Web page,” the author argues that the various publications (including The Wall Street Journal, Forbes, and Gizmodo) characterizing Facebook’s conduct as “a privacy invasion” are incorrect. He writes that there is “a simple reason why”: “That ‘Like’ is only added to the page’s counter. There is no way to tell who added the like . . . [and] [i]f you do not reveal something said or shared in private to others, you are not invading their privacy.” The author goes on to demonstrate the incrementation of the social plugin count, and he explains that “[t]here is no reason for anyone to be upset about Facebook doing this.” He notes that “email providers like Gmail scan user emails all the time. Gmail does it to show relevant ads, fight spam, and slow down viruses.” He also writes that “services across the Internet use whatever method they can to keep track of the popularity of Webpages Google has a list of trends. The New York Times keeps track of most emailed stories.” He concludes that “there is nothing to see here,” telling people to “[m]ove along.” In an update to the article, the author quotes a Facebook spokesperson’s statement on the issue, which noted that “‘[w]hat makes up the number shown on [a] Like button [includes] [t]he number of inbox messages containing this URL as an attachment.’” Page 2 of 22 APP. 40 Exhibit E: Evidence of Implied Consent Source Bates Number Publication Summary 7 FB000000213 Facebook, Update to Messaging and a Test, Facebook (December 20, 2012), available at http://newsroom.fb.com/news/2012/12/upd ate-to-messaging-and-a-test/. In this News Post, Facebook announces updates to Facebook Messenger on mobile and Facebook Messages on the web. The article notes that “[i]n 2010 we introduced the Other folder, where less relevant messages go,” and it continues, “We’ve heard that messages people care about may not always be delivered or may go unseen in the Other folder. As we announced last month, we’re replacing the ‘Who can send me Facebook Messages’ setting with up-front filters that help to address this feedback.” The article then describes the two types of filtering that will be made available (Basic and Strict), noting that “Facebook Messages is designed to get the most relevant messages into your Inbox and put less relevant messages into your Other folder. [Facebook] [relies] on signals about the message to achieve this goal. Some of these signals are social—we use social signals such as friend connections to determine whether a message is likely to be one you want to see in your inbox. Some of these signals are algorithmic—we use algorithms to identify spam and use broader signals from the social graph, such as friend of friend connections or people you may know, to help determine relevance.” The article concludes that Facebook will “continue to iterate and evolve Facebook Messages over the coming months.” 8 FB000000221 Samantha Murphy Kelly, Facebook: We’re Not Liking Brand Pages For You, Mashable (October 4, 2012), available at http://mashable.com/2012/10/04/facebookbrand-like/. In this article, the author notes that “[i]t’s been widely reported on Thursday that Facebook is scanning messages sent to others with attached links to better gauge their interests and add to a brand’s Link count,” writing the information is “used only on the back-end for publishers to see the analytics of articles and shared URLs.” She quotes a Facebook spokesperson’s statement on the issue, and she writes that “[e]mail services such as Gmail have long taken this approach to target its users with ads or fight against viruses based on content written.” She concludes that “Facebook’s developer page related to the Like button states that the number of Likes is derived by the number of likes in the URL and the number of shares. This includes copying and pasting a link back to Facebook. It also includes the number of inbox messages containing the URL as an attachment.” 9 FB000000234 Ryan Singel, Facebook’s E-mail Censorship is Legally Dubious, Experts Say, Wired (May 6, 2009), available at http://www.wired.com/2009/05/facebookse-mail-censorship-is-legally-dubiousexperts-say/. Here, the author reports that “legal experts say Facebook may have gone too far [in] blocking not only links to torrents published publicly on member profile pages, but also examining private messages that might contain them, and blocking those as well.” The author writes that Facebook messages are “governed by the Electronic Communications Privacy Act, which forbids communications providers from intercepting user messages, barring limited exceptions for security and valid legal orders,” opining that “[w]hile the sniffing of e-mails is not unknown—it’s how Google serves up targeted ads in Gmail and how Yahoo filters out viruses, for example—the notion that a legitimate e-mail would be not be delivered based on its content is extraordinary.” The author notes that then-Facebook Chief Privacy Officer Chris Kelly “acknowledged that the site censors user messages based on links [b]ut [ ] insisted that Facebook has the legal right to do so, because it tells users they cannot ‘disseminate spammy, illegal, threatening or harassing content.’” The author quotes Mr. Kelly: “‘Just as many e-mail services do scanning to divert or block spam, prevent fraudulent, unlawful or abusive use of the service—or in the case of some services, to deliver targeted advertising—Facebook has automated systems that have the capability to block links. ECPA expressly allows Facebook to operate these systems. The same automated system that blocks these links may also be deployed where there is a demonstrated disregard for intellectual property rights.’” Page 3 of 22 APP. 41 Exhibit E: Evidence of Implied Consent Source Bates Number Publication Summary 11 FB000000248 Julian Evans, Facebook Launches AntiMalware URL Scanning Service, Julian Evans Blog.com (October 3, 2011), available at https://www.julianevansblog.com/2011/10/ facebook-launches-anti-malware-urlscanning-service html/. Here, the author reports that “Facebook is introducing URL (link scanning) protection for its users as from today (Oct 3rd, 2011).” The author notes that Facebook will “analyze each URL in real-time for potential malicious content.” He acknowledges that “Facebook already scans URLs for malicious links,” noting that “by adding Websenses cloud-based malware technology . . ., they further enhance the security offering to Facebook users.” He concludes that “[o]ne can only applaud Facebook for continuing to build user privacy and protection, even if it is becoming rather more complex for end users to understand.” 12 FB000000146 Jennifer Valentino-DeVries and Ashkan Soltani, How Private Are Your Private Facebook Messages?, Wall Street Journal (October 3, 2012), available at http://blogs.wsj.com/digits/2012/10/03/how private-are-your-private-messages/. In this article, the authors report that an online video shows that Facebook “scans the links you’re sending—registering them as though you ‘Like’ the page you sent,” which the authors characterize as “one example of how online messages that seem private are often actually examined by computers for data.” They note that “[e]mail providers such as Gmail have long reviewed messages in order to spot spam and place ads,” and that Facebook has previously disclosed that the company “analyze[s] messages to filter spam and to detect conversations that could be related to criminal behavior.” The authors also indicate that Facebook’s Developer Guidance discloses that “‘the number of inbox messages containing’ a link to a page will count as ‘Likes’—indicating that the recording of these links isn’t some sort of new bug.” They opine that it is “not clear from Facebook’s data use policy that regular users would expect links in their messages to be scanned this way,” because “the policy simply says generally that Facebook gets ‘data about you whenever you interact with Facebook,’ including when you ‘send or receive a message.’” The authors update their article with a Facebook spokesperson’s statement about the social plugin bug. Additionally, reader comments on the article include a statement by someone who writes that he or she is “[n]ot surprised at all”; another who suggests that “Facebook should update their data use policy to reflect what they already disclose in their guidelines to developers”; and another who states that he is “fine with Google/FB scanning messages as long as private information isn’t released” but does not like “the ability for scammers to inflate the like score with this method.” 13 FB000000281 Joey Tyson, Relevant Ads That Protect Your Privacy, Facebook (September 30, 2012), available at https://www facebook.com/notes/facebookand-privacy/relevant-ads-that-protect-yourprivacy/457827624267125/. In this Facebook Note, the author describes several features that “give advertisers new ways of reaching people who use Facebook.” The author writes that “[m]any sites across the web provide free services by including advertisements. Facebook is no exception, and as we pursue our goal of making the world more open and connected, we have designed our service to show ads that help people discover products that are interesting to them. We also recognize that our users trust us to protect the information they share on Facebook.” The author also writes, “Advertising helps keep Facebook free. We believe we can create value for the people who use our services every day by offering relevant ads that also incorporate industry-leading privacy protections. In our view, this is a win-win situation for marketers and for you.” Page 4 of 22 APP. 42 Exhibit E: Evidence of Implied Consent Source Bates Number Publication Summary 14 FB000000331 KS Sandhya Iyer, Likes Shared Privately Here, the author writes that “Facebook is keeping an eye on your private messages for URLs that have Like buttons and on Facebook Increase Page’s Like Count, should be increased,” noting that “The Next Web further pointed out that the Like button entry on the Facebook NDTV Gadgets (October 5, 2012), Developers page states that the number shown on Like buttons on other websites is a total of likes of that URL, shares available at http://gadgets ndtv.com/social- of that URL, likes and comments on Facebook stories about that URL and inbox messages containing that URL as an networking/news/links-shared-privately-on- attachment.” The author asks, “[H]ow big a deal is it and does it invade Facebook user privacy?” She answers, facebook-increase-pages-like-count“Probably not,” because the “‘Like’ is only added to the page’s counter [and] It does not reveal who added the Like.” 275993/. The article concludes that “[i]f you do not reveal something said or shared in private to others, you are not invading their privacy.” She quotes a Facebook spokerperson’s statement on the issue, and she writes that “Facebook isn’t going to be axed for this move for the simple reason that email providers like Gmail scan user emails all the time” in order to “show relevant ads, fight spam, and slow down viruses.” The author further writes that “[w]hat Facebook is doing is just adopting one of the many services of tracking the popularity of Webpages.” She quotes Facebook’s statement that its “systems parse the URL being shared in order to render the appropriate preview, and to also ensure that the message is not spam.” 15 FB000000372 Doug Mataconis, Your Facebook Chats are Being Monitored, By Facebook, Outside the Beltway (July 13, 2012), available at http://www.outsidethebeltway.com/yourfacebook-chats-are-being-monitored-byfacebook/. The author of this article writes that “Mashable is out with a report that Facebook routinely monitors user chats for suspicious or criminal activity,” quoting several paragraphs of the Mashable article. He notes that the “news was first broken” in a Reuters article from July 2012, which “describes one incident in which [Facebook’s] software did in fact catch a child predator.” The author continues, “As it turns out, this is all covered in the company’s Privacy Policies,” quoting two paragraph’s of Facebook’s Privacy Policy. He concludes that “[i]t’s hard to argue with what Facebook is doing here,” stating that “there are some privacy concerns here, but Facebook is a private company and free to set its own policies on these issues” and that “it has a corporate brand to protect, not to mention the potential of liability, from being known as a place where parents can’t be sure that their teenager children can be safe.” 16 FB000000170 Britney Fitzgerald, New Facebook Bug Scans Messages, Increases ‘Likes’: What You Need To Know, Huffington Post Tech (October 4, 2012), available at http://www huffingtonpost.com/2012/10/04 /new-facebook-bug_n_1940339 html/. In this article, the author opens by noting that “Facebook is experiencing a bug—but it’s not quite the privacy breech [sic] that’s been previously reported by multiple news sources.” She reports that Facebook “has been scanning private messages for links to third-party websites that use Facebook’s ‘Like’ button, a social plug-in that lets users interact with a brand’s products, news articles and other types of content on webpages (without directly visiting the Facebook Page for that brand).” She writes that “Likes also increase when a Facebook user sends another user a message containing a URL to a page featuring the ‘Like’ button; this should only up the ‘Like’ count by one, but it’s actually inflating the count by two.” The author writes that Facebook “insists that this is nothing new,” quoting a Facebook representative, who noted that the issue where “‘counts are jumping by two . . . is a bug [but . . . the actual shares going up when things are sent in messages—that is standard behavior and you can find that in our documentation.’” The author further indicates, “All information posted on the social networking site is accessible for company use. Thus, if you were to share the URL for this article through a Facebook message, Facebook can check out what you’re sending and adjust the ‘Likes’ at the top of this page—whether you clicked ‘Like’ on it or not.” Page 5 of 22 APP. 43 Exhibit E: Evidence of Implied Consent Source Bates Number Publication Summary 17 FB000000177 Michelle Fitzsimmons, Report: Facebook scanning private messages, Liking on users’ behalf, Tech Radar (October 4, 2012), available at http://www.techradar.com/us/news/internet /report-facebook-scanning-privatemessages-liking-on-users-behalf-1101999/. In this article, the author writes that various news sources have indicated that “Facebook is scanning users’ private messages and automatically issuing Likes on their behalf.” The author notes that a Facebook spokesperson directed him to the the “Like Button” section of Facebook’s developer site, and “specifically to information titled ‘What makes up the number shown on my Like button?’.” The author quotes the site, which she writes “explained that ‘the number shown is the sum of’ [items including] [t]he number of inbox messages containing this URL as an attachment.” She concludes that “[m]essage scanning isn’t anything new” and notes that “Gmail scans user emails to create targeted ads and Facebook also reportedly scans user messages to look for sexual predators and child pornography.” 18 FB000000183 Matt Hicks, See the Messages that Matter, Facebook (February 11, 2011), available at https://www facebook.com/notes/facebook/ see-the-messages-thatmatter/452288242130/. In this Facebook Note, the author “announce[s] the next evolution of Messages,” providing detail about the new functionality of the Messages product. He indicates that users may now use SMS, chat, email or Messages; Facebook is providing an “@facebook.com” email address to every person on Facebook who wants one; Facebook messages will take the form of a “single conversation”; and that “your Inbox will only contain messages from your friends and their friends,” and “[a]ll other messages will go into an Other folder where you can look at them separately.” He concludes by inviting readers to take a tour of Messages and asking for thoughts and feedback. 19 FB000000187 Jessica Lee, The Facebook Like Button, Dissected, Bruce Clay (May 25, 2011), available at http://www.bruceclay.com/blog/facebooklike-button/. Here, the author examines the “Like button phenomenon, its various uses and why you need to be liked to survive in the age of online marketing.” She provides an overview of the “two variations” of the “Like” button, writing with respect to the “Like” button social plugin that “Facebook notes that the number of likes shown on any given Web page or object is the sum of [items including] [t]he number of inbox messages containing this URL as an attachment.” 20 FB000000251 In this article, the author discusses in detail how and why Facebook “completedly redesign[ed] its messaging system” in Tekla S. Perry, The Reengineering of Facebook Messages, IEEE Spectrum 2009. The author discusses the rationale behind the “informal formatting” of the Messages product and the decisions (November 2, 2011), available at to remove subject lines and store live chats in the same thread as messages. After explaining how Facebook engineers http://spectrum.ieee.org/computing/softwar determined how to store messages, the author writes that “the engineers turned to the problem of spam, a bane of e-mail e/the-reengineering-of-facebook-messages/. services.” She writes, “While traditional spam filters look mostly at message content, the spam filters built into Facebook messages also pay a lot of attention to who the message senders are. Messages from your friends and friends of friends bypass the spam filters and go directly into your in-box, unless you’ve changed the default or previously moved messages from that person out of your in-box; messages from people you aren’t connected to through a friend, along with announcements from organizations or businesses, go into a folder called ‘other.’ Messages with spamlike content and no friend-of-friend connection go into a separate spam file, the link to which is tucked away at the bottom of the ‘other’ mailbox and requires scrolling past every message in that mailbox to be seen.” The author concludes by opining that “the future of everyday communications will look a lot more like Facebook Messages.” 21 FB000000278 Donna Tam, Facebook Processes More Than 500 TB of Data Daily, Cnet (August 22, 2012), available at http://www.cnet.com/news/facebookprocesses-more-than-500-tb-of-data-daily/. Here, the author writes that “[s]ince Facebook uses [its] data to build its user experience, it wants teams from across the company—whether they sell ads or build functions—to be able to access any of the data as needed.” The author notes that the Facebook employee who “runs Facebook infrastructure” indicated that “this keeps the creation and improvement of Facebook features as fast as possible . . . . [and [t]hese nearly real-time efforts apply to most functions throughout the site because people won’t use the site if the personalized experience is poor, or slow.” The author also writes that Facebook keeps all of its data in one place for “easy access,” but that Facebook “has a zero-tolerance policy when it comes to any abuse from this broad access.” Page 6 of 22 APP. 44 Exhibit E: Evidence of Implied Consent Source Bates Number Publication Summary 22 FB000000285 Sam Biddle, Facebook is Reading Your Messages and Liking Things For You (Updated: Not as Bad as We Thought), Gizmodo (October 4, 2012), available at http://gizmodo.com/5948948/facebook-isreading-your-messages-and-liking-thingsfor-you/. Here, the author writes that Facebook’s “scanning increases the Like count for a given page Like-able link just by you talking about it” in a Facebook message. He indicates that “[a]uto-scanning is nothing new: Gmail has done it since day one to serve us ads,” and he opines that “there are serious potential personal consequences here.” The author continues by noting that “[i]t turns out this was just a very unlikely coincidence that played out in more than one place—the auto-liking only applies to external links with embedded Facebook liking.” He illustrates how a message increments the social plugin count, writing that “your name isn’t being associated publicly with something you’re talking about privately—but if even a mention is enough to kick up a Like, it seems like that’s pretty heavily diluting (even further) what ‘like’ even means—from preference to mere reference.” The author updates the article by quoting a Facebook spokesperson’s statement on the issues raised in the article. 23 FB000000301 Paul Shea, Facebook Confirms Peeking at Private Messages, Value Walk (October 4, 2012), available at http://www.valuewalk.com/2012/10/facebo ok-inc-fb-peeking-at-private-messages/. Here, the author reports that Facebook “confirmed that it scans private messages between users,” and that “these scans [ ] caused the likes of linked content to increase.” He writes that “a series of bots [ ] scan private messages for links to content that contains ‘Like’ buttons [and] [i]f a Like button is detected, however, a bug is activated, whereby the linked content has its likes increased by two.” The author further writes that it “may be news to some long time Facebook users” that the “‘Like’ counter measures not just clicks on a Like button, but takes into account sharing of the content, as well as comments on the content, and now private messages.” He continues, “Emil Protalinski, the writer at thenextweb.com who originally picked up on this story, rightly points out that the scanning of private messages for data on content is not the same as scanning for the same data on comments, or public declarations of ‘Like,”‘ and he notes that “[t]his is not the first time a company has been indicted for scanning the content of users’ private messages,” writing that “Google Inc and other web mail providers have been scanning users’ emails for years, in order to pick the advertisements best suited to them.” He quotes a Facebook spokesperson’s statement on the issue, and he concludes that “Facebook Inc [sic] has certainly not crossed a line with this latest news, any more than they have on hundreds of other occasions,” and that “[b]ecause of the nature of the business, the company will be dodging privacy issues for as long as it operates.” 24 FB000000305 Brittany Darwell, Facebook Clarifies How In this article, the author reports that Facebook “responded to reports today that alleged the social network was Like Plugin Works, Addresses Privacy scanning private messages and Liking pages on users’ behalf” by stating that “the Like count of an article or webpage Concerns, Social Times (October 4, 2012), will increase when users share the link via direct messages [but] that no private information is shared . . . URLs sent available at through private messages are not shown publicly on user profiles and users will not see a friend’s name or photo next http://www.adweek.com/socialtimes/facebo to a Like button if the person shared the article privately.” The author opines that is it not “completely clear to ok-clarifies-how-like-plugin-worksoutsiders that the total includes actions that were made by clicking the button directly, as well as the number of times addresses-privacy-concerns/285167/. the link was copy-pasted into a Facebook post or message, which is why some users thought the social network had a security flaw.” She acknowledges that “Facebook explains this in the FAQ about the Like button plugin,” quoting the relevant FAQ. The author writes that “we see this as similar to site visitor widgets, which increase whenever a user visits a webpage but do not reveal who visited” and that accordingly, “privacy implications are minimal.” Page 7 of 22 APP. 45 Exhibit E: Evidence of Implied Consent Source Bates Number Publication Summary 25 FB000000328 Rick Burgess, Facebook Bug Silently Tallies Up Extra ‘Likes’, TechSpot (October 5, 2012), available at http://www.techspot.com/news/50416facebook-bug-silently-tallies-up-extralikes.html/. This article reports that Facebook is facing “scrutiny” after a “security researcher uncovered a flaw in [Facebook’s] ‘like’ system which appears to be responsible for liking sites an unintended number of times.” The author writes that “Facebook reiterated though that some behaviors will generate likes without explicitly liking something, such as messaging a URL to a friend,” though “some find the sincerity of such a practice nebulous.” He writes that “[]on Facebook for Developers . . . there are actually four ways to generate likes [and] . . . only one of those methods actually requires users to click a like button.” He concludes by citing Facebook’s recent “initiative which intends to fortify the integrity of Facebook’s likes and shares,” including deploying “automated tools . . . with the intent of deleting disingenuous likes that were determined to be purchased or originate from malware or compromised accounts.” 26 FB000000376 Lisa Vaas, How Facebook Catches WouldBe Child Molesters by Analyzing Relationships and Chat Content, Naked Security (July 16, 2012), available at https://nakedsecurity.sophos.com/2012/07/ 16/facebook-child-molester/. In this article, the author reports that “[]law enforcement is hailing Facebook for using its little-known data monitoring technology to spot a suspicious conversation about sex between a man in his early thirties and a 13-year-old girl from Florida,” and that “Facebook doesn’t talk much about this technology, which scans postings and chats for criminal activity.” She notes that Facebook’s then-Chief Security Officer Joe Sullivan said that Facebook’s “monitoring software analyzes relationships to find suspicious conversations between unlikely pairings” and “relies on archives of real-life chats that preceded sexual assaults.” She notes that Mr. Sullivan said that “‘the last thing the company wants is for its users to feel like they’re being eavesdropped on.’” The author ends by quoting Mr. Sullivan, who says, “‘We’ve never wanted to set up an environment where we have employees looking at private communications, so it’s really important that we use technology that has a very low false-positive rate.’” 27 FB000000391 Here, the author writes that “[w]hether you realize it or not, a bundle of sophisticated technology is constantly scanning Adam Estes, Facebook’s Spying on You For a Good Cause, Motherboard (July 13, through Facebook interactions—wall posts, messages, chats—looking for sexual predators.” The author reports that 2012), available at Facboook’s program is “part of an aggressive effort the social network has made over the past few years to protect the http://motherboard.vice.com/blog/facebook- safety of its 13- to 18-year-old users, and few would argue that the stated goals of the program aren’t sound.” He s-reading-your-messages-but-it-s-for-awrites that “there’s something unnerving about Facebook reading your messages,” because “[p]reventing crime is one good-cause/. thing, but surveilling the most intimate user behavior is something completely different.” The author also notes that “Facebook is obviously aware of the privacy concerns and insist that their technology only spots the bad guys,” quoting Facebook’s then-Chief Security Officer Joe Sullivan, who says, “‘We’ve never wanted to set up an environment where we have employees looking at private communications, so it’s really important that we use technology that has a very low false-positive rate.’” He concludes that “it’s easy to think of Facebook’s anti-pedophile software as just another form of moderation” because “[t]he vast majority of the scanning is also algorithmic, so it’s not like you have a bunch of Facebook employees poring over your every word,” writing that “[i]n truth, it’s a machine that’s trying to spot patterns and red flags” for “a good cause.” Page 8 of 22 APP. 46 Exhibit E: Evidence of Implied Consent Source Bates Number Publication 28 FB000000393 Cale Guthrie Weissman, Facebook Uses Scanning Technologies, Alters Authorities About Content, OpenNet Initiative (July 16, 2012), available at https://opennet net/blog/2012/07/facebookuses-scanning-technologies-alertsauthorites-about-content/. 29 FB000000395 30 FB000000399 Summary Here, the author writes that “Facebook regularly scans user content for criminal activity, but the monitoring program is something the social media giant has generally kept quiet about. “ Quoting a report from Reuters, the author acknowledges that “this sort of scanning is commonplace for platforms like Facebook—most large social media companies scan chats for inappropriate language and exchange of personal information,” but he opines that many social media platforms “walk a tightrope between utilizing these tactics to safeguard against illegal activity and providing a less restrictive social media platform that will engage users.” The author writes that “[s]canning users’ content is not new terrain for Facebook,” citing information leaked that April [2012?] “showing the kinds of user information Facebook releases to authorities when subpoenaed” and the “Information for Law Enforcement Authorities” section of Facebook’s website. The author quotes Facebook’s then-Chief Security Officer Joe Sullivan, who says, “‘We’ve never wanted to set up an environment where we have employees looking at private communications, so it’s really important that we use technology that has a very low false-positive rate.’” He concludes that “[s]canning technologies of this design are just beginning to come to the forefront for various websites,” and that Facebook’s “example highlights the tactics used and suggests a possible upward trend in surveillance by social media platforms.” The author of this article writes that Facebook “uses monitoring software that can scan and flag suspicious messages to Michael Walsh, Did You Know that Facebook Monitors Postings and Chats for minors from potential predators.” Citing an instance in which Facebook’s software effectively flagged “suspicious Sexual Predators?, NY Daily News (July conversations between a man in his early thirties and a 13-year old girl,” the author opines that Facebook’s 16, 2012), available at “surveillance practice is fraught with legal complexity” and “Facebook tends to avoid comment on this practice, http://www nydailynews.com/news/crime/f because the organization doesn’t want to create scare stories or stir surveillance paranoia.” The author quotes acebook-monitors-postings-chats-sexual- Facebook’s then-Chief Security Officer Joe Sullivan, who says, “‘We’ve never wanted to set up an environment where predators-article-1.1115392/. we have employees looking at private communications, so it’s really important that we use technology that has a very low false-positive rate.’” The author concludes that “[t]o minimize the risk of inappropriate surveillance the software and procedures are designed to err on the side of monitoring caution.” Jemima Kiss, Facebook Puts Faith in its Software Smarts to See Off Sexual Predators, The Guardian (April 15, 2010), available at http://www.theguardian.com/technology/20 10/apr/16/facebook-software-sexualpredators/. This article reports that “Facebook has developed sophisticated algorithms to monitor its users and detect inappropriate and predatory behaviour, bolstering its latest raft of initiatives to improve the safety of its users.” The author writes that “[h]aving launched an education campaign, an improved reporting procedure and a 24/7 police hotline,” Facebook indicated that it “has introduced a number of algorithms that track the behaviour of its users and flag up suspicious activity.” The author also notes that Facebook has another filter, “common on web publishing sites [ ] [that] scans photo uploads for skin tones and blocks problem images.” She quotes Facebook employee Matt Kelly, who says that Facebook “balance[s] its duty to respect its users while meeting its legal obligations,” and that Facebook’s “corporate philosophy about data is that the user is in control, and they choose how to share and distribute it. She ends by noting that one privacy campaigner believes Facebook “needed to do more to stop persistent stalkers and bullies who could use multiple identities, and cautioned against the automated profiling of users.” Page 9 of 22 APP. 47 Exhibit E: Evidence of Implied Consent Source Bates Number Publication Summary 31 FB000000420 Josh Wolford, Should Facebook Monitor Chats to Help Snag Child Predators?, WebProNews (August 23, 2012), available at http://www.webpronews.com/rememberhow-facebook-is-monitoring-chats-forcriminal-activity-well-it-worked-kind-of2012-08/. Here, the author begins by stating that “social networks are social – you’re actively sharing content with the world,” and “anybody who thinks they can maintain a pristine level of privacy and security while still enjoying the benefits of a social community is probably deluding themselves.” He writes that “recently, it was revealed that Facebook actively patrols user communications for unlawful activities,” and that “Facebook is actively monitoring our chats and messages.” The author notes that Facebook “revealed that it’s common practice for their teams to scan chats, searching for criminal activity” and that “[i]t’s mostly algorithms that handle this part, but once something is flagged Facebook employees make the final decision on whether or not it merits calling the authorities.” He posits that “[t]here’s really no denying that it can work,” as “[s]canning chats for suspicious activity can help to thwart child predation,” noting that ‘there are still privacy concerns to consider.” He writes that “[n]ot everyone is convinced that Facebook has the right to monitor ‘private’ communications,” but states, “Then again, you are using their (free) service to send and receive communications, and at least now it’s with the public knowledge that the company may be monitoring them.” He adds that Facebook is “not the only one[ ] engaging in this type of monitoring.” 32 FB000000066 Sue Keogh, How accurate is the Facebook Like count?, The Wall (April 8, 2011), available at http://wallblog.co.uk/2011/04/08/howaccurate-is-the-facebook-like-count/. Referencing a blog post that suggests that the “accuracy” of the number of “Likes” indicated for a given Facebook page on the website may be as low as 39%, the author of this article writes that the “Like” count on a third-party website also reflects additional actions other than affirmative clicks on the “Like” button. The author quotes Facebook’s Developer Guidance, which explains that “‘[t]he number shown is the sum of . . . [actions including] [t]he number of inbox messages containing this URL as an attachment.’” She concludes that “the information is out there, if you know where to look. I might even click Like to help spread the word.” 33 FB000000068 Bianca Bosker, Facebook’s Paid Messages Test Taxes You For Being Social, Huffington Post Tech (December 20, 2012), available at http://www huffingtonpost.com/biancabosker/facebook-messages-test_b_2341521 html/. This article concerns Facebook’s testing of a new program that permits individuals with whom a user is connected to “pay to re-route their message from the ‘other’ heap straight to your inbox.” The author notes that this feature “is being rolled out in conjunction with new filters for Facebook’s messaging system that aim to ensure important messages don’t go unseen in the ‘Other’ inbox.” She also writes that “[a]s a user who receives no shortage of spam messages, I’m all for cutting back on clutter or fining advertisers who want to get hold of me in my inbox, uninvited,” but she labels the payment proposal “anti-social.” The author quotes Facebook’s statement explaining that the test is “‘designed to address situations where neither social nor algorithmic signals are sufficient. For example, if you want to send a message to someone you heard speak at an event but are not friends with, or if you want to message someone about a job opportunity, you can use this feature to reach their Inbox. For the receiver, this test allows them to hear from people who have an important message to send them.’” Acknowledging that “[k]eeping savvy scammers at bay is a gargantuan challenge for Facebook,” she argues that “Facebook profits from allowing people—and, most likely, brands—to take up your time when you made clear you didn’t want them to,” and she expresses concerns that the test is “more like a fix for Facebook’s profit push than a solution to overcrowded inboxes.” 34 FB000000073 How to get share counts using graph API, Stack Overflow (June 15, 2012), available at http://stackoverflow.com/questions/569927 0/how-to-get-share-counts-using-graphapi/. On this developer forum, a user requests “a way to get share counts of an URL using graph API.” A user named “Jim Rubenstein” responds on June 15, 2012, clarifying that Facebook’s Graph API is a way to get the “share count,” which is “not equal to the one you see on the Like button, since that number is the sum of [items including] [t]he number of inbox messages containing [the] URL as an attachment.” Page 10 of 22 APP. 48 Exhibit E: Evidence of Implied Consent Source Bates Number Publication Summary 35 FB000000077 Jessican Guynn, Facebook Looks To Cash In this article, the author notes that “[p]rofiles, status updates and messages all include a mother lode of voluntarily In On User Data, Los Angeles Times (April provided information,” and that Facebook uses this information “to help advertisers find exactly who they want to 17, 2011), available at reach.” The author describes how Facebook is helping advertisers reach their target audiences with precision: http://articles.latimes.com/2011/apr/17/busi “Facebook doesn’t have to guess who its users are or what they like. Facebook knows, because members volunteer this ness/la-fi-facebook-ads-20110417/. information freely—and frequently—in their profiles, status updates, wall posts, messages and ‘likes.’ It’s now tracking this activity, shooting online ads to users based on their demographics, interests, even what they say to friends on the site—sometimes within minutes of them typing a key word or phrase.” The author also suggests that Facebook’s ability to “mine data and sell advertising based on what its members voluntarily share amounts to electronic eavesdropping on personal updates, posts and messages that many users intended to share only with friends,” but she concedes that “any information users post on the site—hobbies, status updates, wall posts—is fair game for ad targeting.” The article concludes by quoting a Facebook user who says she enjoys “receiving ads from merchants”: “‘I don’t feel any weird privacy thing. We are all putting everything out there already.’” 36 FB000000083 Rove Monteux, Facebook Graph API In this post, the blogger explains that “simply private messaging an URL to some people on Facebook will increase the exploit that let[s] you pump up to 1800 ‘like’ count for the given URL.” ‘Likes’ in an hour, Rove Monteux (October 5, 2012), available at http://rmonteux.wordpress.com/2012/10/05 /facebook-graph-api-exploit-that-lets-youpump-up-to-1800-likes-in-an-hour/. 37 FB000000087 Bianca Bosker, Facebook Home’s Ultimate Goal: Ingesting Your Messages, Huffington Post Tech (April 11, 2013), available at http://www huffingtonpost.com/2013/04/11 /facebook-homemessages_n_3063609 html/. Reporting on the launch of Facebook Home, Facebook’s smartphone software, the author writes that “tech industry analysts note that the sooner people channel their chatting through Facebook, the sooner Facebook can turn messaging from communication between friends into a moneymaker that involves brands. More messaging will give Facebook more data it may use to provide advertisers with personal, personalized ways of interacting with its members.” The author notes that “Facebook’s Messenger app currently gives people the option to attach location information to each post,” and that Facebook next “might analyze the content of messages to serve up ads targeted to each conversation, much like Gmail.” She quotes an analyst who says, “‘[Facebook’s messaging platform] is not just for connecting people. It’s for connecting brands, too.’” The author also suggests that “[w]hether users would accept advertising via messages—private, personal and traditionally off-limits to brands—remains to be seen,” quoting an analyst who says that “‘[t]here’s a certain creepy factor to that, but users are getting much more comfortable trading privacy for convenience.’” Page 11 of 22 APP. 49 Exhibit E: Evidence of Implied Consent Source Bates Number Publication Summary 38 FB000000097 Emil Protalinski, Facebook Confirms It Is Scanning Your Private Message For Links To Increase Like Counters, The Next Web (October 4, 2012), available at http://thenextweb.com/facebook/2012/10/0 4/facebook-confirms-it-is-scanning-yourprivate-messages-for-links-so-it-canincrease-like-counters/. Here, the author writes that Facebook “is monitoring your private messages for links that have Like buttons and should be increased,” and he quotes a Facebook spokerperson’s statement about the issue: “We did recently find a bug within our social plugins where at times the count for the Share or Like goes up by two, and we are working on [a] fix to solve the issue now. To be clear, this only affects social plugins off of Facebook and is not related to Facebook Page likes. This bug does not impact the user experience with messages or what appears on their timelines.” The author indicates that while this was “news” to him, “this was clearly the case before as on the Like button web page over on Facebook Developers, the social networking giant says the number shown on a Like button is the sum of [items including] [t]he number of inbox messages containing this URL as an attachment.” The author writes that he had “known for a while that the Like button isn’t a counter of just Likes,” but he opines that “private messages . . . have privacy questions attached to them.” He notes that Facebook confirmed that “[w]hen the count is increased via shares over private messages, no user information is exchanged, and privacy settings of content are unaffected. Links shared through messages do not affect the Like count on Facebook Pages.” In an update to the article, the author quotes Facebook’s explanation of the situation: “‘Our systems parse the URL being shared in order to render the appropriate preview, and to also ensure that the message is not spam.’” 39 FB000000104 Alicia Eler, Facebook Is Using Your Data Whether You Like It Or Not, Read Write (November 28, 2011), available at http://readwrite.com/2011/11/28/facebook_ is_using_your_data_whether_you_like_it_ or/. Here, the author discusses the implications a European Commission Directive that purportedly “ban[s] targeted advertising unless users specifically say they want it” may have on Facebook. In response to the question, “Why isn’t this happening in America?,” the author writes, “[Because] [a]ll 800 million Facebook users agree to let the company use their personal information.” She further writes that “Facebook has information about a user’s friends, family, education background in addition to Facebook such as ‘likes’ and everything that gets posted to Facebook Walls . . . Messages and ‘chats’ are stored, too, even if the user deletes them.” She also highlights that “Facebook denies tracking peoples’ behavior to serve advertising . . . denies selling users’ personal information to third parties . . . claims that advertisers only see ‘anonymous and aggregate information,’ using that to serve up targeted ads . . . [and] does not target individual users.” The author also notes that “with an IPO in the works, and Facebook’s move toward becoming self-sufficient, there’s no denying that advertising on the site has increased,” concluding by noting recent Federal Trade Commission scrutiny of Facebook’s practices. 40 FB000000106 Facebook Launches Messaging System With In-Bound Message Filter, Out Law (November 10, 2010), available at http://www.out-law.com/page-11555/. This article reports on Facebook’s launch of the Messages product and its features. The author quotes Facebook’s statement that “‘[i]t seems wrong that an email message from your best friend gets sandwiched between a bill and a bank statement . . . It’s not that those other messages aren’t important, but one of them is more meaningful. With new Messages, your Inbox will only contain messages from your friends and their friends. All other messages will go into an Other folder where you can look at them separately.’” Noting that “[c]ritics of Facebook’s past approaches to user privacy will be assessing the system closely,” the author quotes a researcher who states that “‘Facebook announced it will not utilize the content of users’ personal messages to target advertising . . . This is surprising, considering doing so is typical among web-based email clients; both Gmail and Yahoo Mail scan users’ messages for keywords in order to better serve relevant advertising.’” The author concludes by quoting a technology consultant who opines that “‘Facebook would be able to track vital information through the Messages system.’” Page 12 of 22 APP. 50 Exhibit E: Evidence of Implied Consent Source Bates Number Publication Summary 41 FB000000109 Ellis Hamburger, How To Find All The Private Messages Facebook Is Hiding From You, Business Insider (December 12, 2011), available at http://www.businessinsider.com/facebookmessages-other-folder-2011-12/. This article explains how to access the “Other Messages” folder in Facebook Messages, noting that “Facebook often groups messages from people you aren’t ‘Friends’ with into a spam-box-evoking ‘Other Messages’ folder.” The author writes, “While most messages that get filtered into your Other folder are in fact spam, it’s definitely worth digging through once in a while.” 42 FB000000110 The author writes that Facebook “scans [] private messages to friends, and when it sees a link to a ‘Likeable’ page, it Kashmir Hill, Facebook Scans Private Messages To Hand Out Public ‘Likes’, doles out ‘Likes’ accordingly.” The article summarizes the findings of the Wall Street Journal article on this topic Forbes (October 4, 2012), available at written by Ashkan Soltani and states, “We already know that Google scans our Gmail to target us with ads and that http://www forbes.com/sites/kashmirhill/20 Facebook scans our inboxes looking for sexual predators and child porn, but many users may not have realized that the 12/10/04/facebook-scans-private-messages- links they were exchanging privately resulted in a public—though anonymous—endorsement of said links.” The to-hand-out-public-likes/. author includes a statement by a Facebook spokesperson on the issue: “‘Many websites that use Facebook’s “Like”, “Recommend”, or “Share” buttons also carry a counter next to them. This counter reflects the number of times people have clicked those buttons and also the number of times people have shared that page’s link on Facebook. When the count is increased via shares over private messages, no user information is exchanged, and the privacy settings of content are unaffected. Links shared through messages do not affect the Like count on Facebook Pages.’” The article also includes an update that reads, “A spokesperson also emphasizes that this [is] a third-party Social Plug-in version of the ‘Like’ button and that it is meant to reflect engagement rather than endorsement.” 43 FB000000119 Ken Yeung, Facebook Updates Messages Feature With Filtering, Tests A Service To Let People Pay To Send Them, The Next Web (December 20, 2012), available at http://thenextweb.com/facebook/2012/12/2 0/facebook-messages-now-with-filters/. Here, the author reports on Facebook’s addition of two “filters to help solve the problem of finding messages that should be seen and seeing those that shouldn’t” to the Messages product. The author notes that “[t]here are many signals that go into determining what gets through to a user’s Message inbox, but one new signal Facebook is testing is allowing some users the ability to pay in order to help get their messages delivered, regardless of friendship status on the network.” He mentions the “Other” folder in Facebook Messages that “basically acts as the catch-all for all communications that it deems to have low relevancy,” positing that “[i]t’s too bad that a few innocent messages get tagged and removed because of this filter.” He concludes by explaining that the settings for the new filters will be “visible right with the Messages screen.” 44 FB000000148 Doug Gross, How You Help Facebook Make Millions, CNN Tech (May 16, 2012), available at http://www.cnn.com/2012/05/16/tech/socia l-media/facebook-users-ads/. In this article, the author writes that the following are the “building blocks” of Facebook as “a multibillion-dollar company”: “Every post you ‘like.’ Every friend you add or fan page you join. Every place you check in, and every Web page you recommend.” The author writes that “Facebook’s unprecedented advertising advantage is built upon the service it provides. As users interact with the site, they gradually build a fuller and fuller picture of themselves. That, in turn, lets Facebook sell advertisers on its ability to put their product in front of the people most likely to be interested.” He notes that Facebook’s advertising model has “made some folks antsy,” quoting an Associated Press/CNBC poll that found that “three out of five users say they have little or no faith that the company will protect their personal information.” The author concludes by quoting an analyst who “‘expects’” that Facebook’s “‘datadriven model [will] keep making money well into the future.’” Page 13 of 22 APP. 51 Exhibit E: Evidence of Implied Consent Source Bates Number Publication Summary 45 FB000000180 Phil Villarreal, Rescue Messages From Facebook’s De Facto Spam Filter, Consumerist (December 11, 2011), available at http://consumerist.com/2011/12/15/rescuemessages-from-facebooks-de-facto-spamfilter/. In this article, the author discusses Facebook’s “Other” folder in the Messages product, writing that “[w]hen Facebook thinks you don’t particularly want to read a message that’s sent your way, it redirects it into” that folder. He writes that “[s]ome users forget to check the box regularly, and others may not even be aware that they have it,” and he provides information about how to access the folder. 46 FB000000194 Greg Finn, The Formula Behind The Facebook ‘Like’ Number, Marketing Land (October 17, 2012), available at http://marketingland.com/the-formulabehind-the-facebook-like-number-24069/. Here, the author reports on the “uproar” surrounding reports that “messages were not only being crawled, but also used towards the overall ‘Like data’ of a page.” He writes that “[o]ne of the important lessons that marketers learned from the situation was that the “‘Like count’ wasn’t really about likes, rather other interactions (and messaging) that occurs on Facebook.” He examines what goes into “Facebook Like data,” quoting Facebook’s Developer site and noting that “four different variables make up the Like number [including] [t]he number of inbox messages containing this URL as an attachment.” He writes that a Facebook spokesperson confirmed that the listed elements are “all counted into the overall Facebook Like data.” 47 FB000000217 JVG, Facebook Tweaks Messages With Inbox Filters and Tests Pay-to-Deliver Option, Venture Beat (December 20, 2012), available at http://venturebeat.com/2012/12/20/faceboo k-messages-inbox/. Here, the author reports on Facebook’s announcement that it was “tweaking the Facebook Messages inbox experience to better ensure that relevant messages get to your inbox.” The author writes that “[t]he new inbox filters, which are rolling out globally, are Facebook’s way of correcting a broken system. The company currently routs messages to your inbox or other folder based on your settings, but it has found that it pushes too many ‘high signal’ messages (read as: messages you probably want to go to your Other folder, where they likely go unseen. To fix the problem, Facebook has created two filters, basic and strict, that will allow certain types of messages to reach your inbox that otherwise would not . . . Both options, however, use the ‘mostly’ terminology to allow for instances when Facebook puts its algorithmic magic to work and determines that a high signal message should get through.” The author also reports that Facebook is lauching a “small test” to allow people that are not your Facebook friends to pay a fee to send messages directly to your inbox. 48 FB000000237 Oliver Chian, Facebook Messages Isn’t a In this article, the author reports on the launch of the Facebook Messages product, arguing that it’s a “problem” that the Gmail Killer—And That’s the Problem, product is not a “Gmail killer.” He provides a “quick rundown” of the features of Facebook Messages, writing that the Forbes (November 15, 2010), available at product “takes all these bits of conversation and keeps a collective conversation history” and “filters messages based http://www forbes.com/sites/oliverchiang/2 on your Facebook social graph.” He notes, for example, that “[t]raditional spam emails will end up in Junk” and that 010/11/15/facebook-messages-isnt-a-gmail- Facebook users “will have control over this filtering too” because they can “move it manually [and] [t]he system will killer-and-thats-the-problem/. remember your choices and change your filtering preferences accordingly.” The author writes that the “main problem” about Facebook Messages is that it “has no way of telling if your mother is sending an important email about your cousin’s wedding, or a forwarded email about the billionth funny cat video ever made.” 49 FB000000276 Jay Yarow, How to Find Facebook Messages that Facebook is Hiding From You, Business Insider (June 15, 2012), available at http://www.businessinsider.com/facebookmessages-spam-filter-2012-6/. Here, the author writes that Facebook “has a pretty aggressive spam filter to keep messages from creepers, and trolls from hitting your Facebook message inbox,” opining that the spam filter is “sometimes too aggressive, and grabs messages that aren’t spam.” The author also provides instructions about how to check Facebook messages in the “Other” folder. Page 14 of 22 APP. 52 Exhibit E: Evidence of Implied Consent Source Bates Number Publication Summary 50 FB000000312 Donna Tam, To Facebook, A Shared Link is As Good as a Like, Cnet (October 4, 2012), available at http://www.cnet.com/news/to-facebook-ashared-link-is-as-good-as-a-like/. Here, the author reports that “[a] recent bug adds two Likes to the count instead of one, and Facebook said it’s working to fix that.” The author writes that “the feature may rankle some users who don’t want to be part of an overinflated counter” or others who “feel violated” that Facebook “know[s] what you’re sending a Friend.” She notes Facebook says “this feature doesn’t affect Like counts on Facebook pages and it’s not an invasion of privacy since the Likes on the plugs-ins are anonymous,” quoting a Facebook spokerperson’s statement on the issue in full. She writes that a Facebook representative also confirmed that “Facebook does scan any links that pass through the network to look for spam,” including “links you send to a friend through messages,” and that “Facebook automatically reviews the links before generating a link preview.” The author concludes by noting that “even if you were sharing a link to show a friend something you don’t like, you’d still be adding to the page’s Like numbers.” 51 FB000000314 Mark Langshaw, Facebook Adding Likes This author reports that a “security expert discovered that sending a web address to a friend automatically adds two on Users’ Behalf, Says Report , Digital Spy likes to that page, suggesting that [Facebook] is scanning private messages.” The author notes that “Facebook has (October 5, 2012), available at responded to the report’s findings and issued a statement denying that privacy information has been exposed,” quoting http://www.digitalspy.com/tech/news/a410 certain portions of Facebook’s official statement, including the following: “Absolutely no private information has been 470/facebook-adding-likes-on-users-behalf- exposed . . . . Each time a person Shares a URL to Facebook, including through messages, the number of Shares says-report html#~p6spSJD5SKvz21/. displayed on the social plugin for that website increases. Our systems parse the URL being shared in order to render the appropriate preview, and to also ensure that the message is not spam. These counts do not affect the privacy settings of content, and URLs shared through private messages are not attributed publicly with user profiles.” The author concludes that “[a]lthough Facebook also stressed that the additional likes are anonymous and will not appear on users’ timelines, critics have pointed out that people who share pages to highlight negative content are making the site appear more popular.” 52 FB000000335 Liz Klimas, Why is the Privacy of Personal Facebook Messages Being Called Into Question—Again?, The Blaze (October 5, 2012), available at http://www.theblaze.com/stories/2012/10/0 5/why-is-privacy-of-personal-facebookmessages-being-called-into-questionagain/. This author writes that the other publications have reported that “including links in private messages [on Facebook]—if these links had a ‘like’ button associated with them –would increase the “likes” on that actual page by two.” The author writes that “Facebook has responded saying that there was a bug identified in the system that was accidentally counting one ‘like’ or ‘share’ of a link or post as two” and was “working to fix it.” She writes that “Facebook emphasized that ‘no private information has been exposed,’” and that this means that “if you receive a message containing a link that has a ‘like’ button, you are not automatically ‘liking’ this item on your Timeline.” The author continues by noting that “it’s probably news for many” that the social plugin counter “is not just measuring clicks but sharing content as well,” and she quotes another publication that reported that “‘[O]n the Like button Web page over on Facebook Developers, the social networking giant says the number shown on a Like button is the sum of: [t]he number of likes of this URL; [t]he number of shares of this URL (this includes copy/pasting a link back to Facebook); [t]he number of likes and comments on stories on Facebook about this URL; [t]he number of inbox messages containing this URL as an attachment.’” The author concludes by quoting a reporter who “writes that Facebook has ‘not crossed a line with this latest news, any more than they have on hundreds of other occasions.’” Page 15 of 22 APP. 53 Exhibit E: Evidence of Implied Consent Source Bates Number Publication Summary 53 FB000000347 Greg Finn, Your Private Facebook Messages Aren’t So Private: Shared Links Count Toward ‘Like’ Data, Marketing Land (October 8, 2012), available at http://marketingland.com/your-privatefacebook-messages-arent-so-private-aslinks-count-towards-like-data-23400/. Here, the author reports that Facebook’s “Like button is an aggregate score from a variety of Facebook actions, including links shared within private messages.” He continues, “Facebook did confirm that the issue of double counts was a bug, but did also confirm that shared messages do count towards the overall ‘like’ data,” and he quotes the Facebook Developer site that “clearly states the following about Like buttons”: “‘The number shown is the sum of: [t]he number of likes of this URL; [t]he number of shares of this URL (this includes copy/pasting a link back to Facebook); [t]he number of likes and comments on stories on Facebook about this URL; [t]he number of inbox messages containing this URL as an attachment.’” He reports that “[t]he fact that private shares gave an endorsement (even if an anonymous one) drew a bit of an uproar” because users sharing a link of a product they don’t like “will still be counted as a ‘like.’” He concludes by quoting Facebook’s statement on the issue. 54 FB000000361 Louis Goddard, Facebook Analyzes Relationships and Chats to Flag Up Sexual Predators, The Verge (July 13, 2012), available at http://www.theverge.com/2012/7/13/31564 99/facebook-paedophile-scanning/. In this article, the author reports that “Facebook automatically scans posts and chat logs for criminal activity, using big data processing techniques similar to those used in targeting advertising to determine the most vulnerable users.” He writes that Facebook’s “scanning tools use factors such as mutual friends, past interaction, distance and age difference—alongside simple phrase searches—to flag potentially nefarious conversations for human moderators” and “rely on archives of previous conversations that are known to have led to sexual assaults, identifying patterns and searching for similar ones.” The author quotes Facebook’s then-Chief Security Officer Joe Sullivan, who says, “‘We’ve never wanted to set up an environment where we have employees looking at private communications, so it’s really important that we use technology that has a very low false-positive rate.’” The author writes, “Privacy issues aside, it would be practically impossible for human moderators to effectively trawl through the vast amount of data generated by more than 900 million users each day.” 55 FB000000388 Fahmida Y. Rashid, Facebook Scans Chats Here, the author reports that “Facebook has technology in place to monitor user conversations for suspicious activity for Criminal Activity, PC Mag (July 13, and notify police when necessary.” She writes that Facebook’s “scanning technology monitors chats for words or 2012), available at phrases that may signal that something is wrong, such as personal information being exchanged or explicit language http://securitywatch.pcmag.com/security/30 being used.” The author writes that Facebook stated, “‘We’ve never wanted to set up an environment where we have 0288-facebook-scans-chats-for-criminal- employees looking at private communications, so it’s really important that we use technology that has a very low falseactivity/. positive rate,’” and she notes that “Facebook security employees don’t see any of the conversations until the scanning technology actually flags the exchange.” She reports that one security advisor found the news to be “‘scary and more than a bit surprising,’” but the author opines that “it’s nice to know that Facebook is keeping a distant eye on chat logs for criminal behavior.” She provides “tips” for people to protect themselves on Facebook and other social networking sites, and she writes that “Facebook relying on software to pre-scan chats protects the company from privacy concerns that someone is monitoring all conversations.” 56 FB000000397 Walter Pacheco, Facebook Scans Conversations For Criminal Activity, Orlando Sentinel (July 13, 2012), available at http://www.orlandosentinel.com/business/t echnology/os-facebook-scans-criminalactivity-20120713-post html/. Here, the author cites a Reuters report stating that “Facebook is scanning users’ chats and posts for possible criminal activity.” The author writes that Facebook’s then-Chief Security Officer Joe Sullivan stated that “Facebook monitors conversations for words and phrases that suggest potential criminal activity, as well as the exchange of personal information between users with a wide age gap,” and that Facebook software “‘searches for words often found in the chat records of convicted criminals, including sex offenders, who used social media to find their victims.’” Page 16 of 22 APP. 54 Exhibit E: Evidence of Implied Consent Source Bates Number Publication Summary 57 FB000000405 Justin Reynolds, Facebook Monitors Potentially Illegal Posts, Chats, WestonReddington-Easton Patch (July 13, 2012), available at http://patch.com/connecticut/westonct/facebook-eyes-posts-about-illegalactivity/. Here, the author reports that, “[u]sing data recognition software, Facebook employees monitor certain users’ posts and chats, scanning them for potentially illegal activity which in some cases has led the social media giant to contact police.” The author writes that CNET reports that Facebook “isn’t actively monitoring all communications on Facebook, as it wants its users to maintain their privacy” and that “[t]he software the company uses to analyze communications which are potentially illegal has a low false-positive rate.” The author concludes by noting that Facebook’s “scanning program looks for certain phrases found in previously obtained chat records from criminals, including sexual predators,” and that “[t]he relationship analysis and phrase material have to add up before a Facebook employee actually looks at communications and makes the final decision of whether to ping the authorities.” 58 FB000000416 This article reports that “[w]hat you say in your private chats and messages on Facebook may not be as private as you Jillian Ryan, Your Facebook Chats are Being Monitored, Find out Why: The think,” citing a Reuters report that indicated that Facebook “employs a mums-the-word technology that scans posts and Social Media Privacy Report, The Private chats for criminal activity.” The author notes that Facebook’s “monitoring came to light” earlier that year when WiFi Blog (July 20, 2012), available at Facebook’s software detected an alleged sexual predator. She quotes Reuters’s description of Facebook’s efforts for http://www.privatewifi.com/your-facebook- detecting criminal activity, which “‘generally start with automated screening for inappropriate language and exchanges chats-are-being-monitored-find-out-why- of personal information, and extend to using the records of convicted pedophiles’ online chats to teach the software the-social-media-privacy-report/. what to seek out.’” 59 FB000000418 Lee Bell, Facebook Scans Private Chats and Posts for Criminal Activity, The Inquirier (July 13, 2012), available at http://www.theinquirer net/inquirer/news/2 191599/facebook-scans-private-chats-andposts-for-criminal-activity/. This article reports that “Facebook scans private chats and posts for criminal activity,” noting that Facebook’s software for “‘scanning postings and chats’” detected an alleged sexual predator. The article quotes a security consultant who states, “‘It shouldn’t surprise anybody that Facebook is trying to make its site a safer place by monitoring for illegal and suspicious behaviour which might bring it into disrepute. Obviously we have to hope that Facebook acts responsibly, and puts measures in place to prevent inappropriate monitoring - or risk a backlash from users.’” 60 FB000000423 Kate Tummarello, Facebook Knows When You’re Chatting About Your Illegal Activities, DCInno (July 13, 2012), available at http://dcinno.streetwise.co/2012/07/13/face book-knows-when-youre-chatting-aboutyour-illegal-activities/. In this article, this author writes that “[a]ccording to online reports, Facebook uses a software that screens private chats to determine if participants are discussing illegal activities.” She quotes a Mashable article on this topic that explains how Facebook’s software works, and she notes that Facebook commented on the issue as follows: “‘We’ve never wanted to set up an environment where we have employees looking at private communications, so it’s really important that we use technology that has a very low false-positive rate.’” She ends by noting that “[b]y keeping most of the private chat records away from employees, Facebook is protecting itself from some privacy advocates.” 61 FB000000358 Here, the author opens by asking, “Ever wonder if Facebook is reading your posts?” He answers, “Well, it is—or, its Will Oremus, Facebook Monitors Your Posts and Chats to Catch Sexual Predators, computers are, at least,” quoting a Reuters article that recounted a case in which Facebook’s software detected a man in Slate (July 17, 2012), available at his thirties allegedly trying to set up a meeting with a 13-year-old girl for sex. The author writes that “[]in Facebook’s http://www.slate.com/blogs/future_tense/20 case, the scanning hasn’t stirred outrage—probably because it seems to be focused on catching sexual predators.” He 12/07/17/online_privacy_facebook_monito concludes by writing, “It seems clear that this technology has the potential to do some good. But that shouldn’t blind rs_your_posts_chats_to_catch_sexual_pred us to the fact that it represents a further erosion of our online privacy, one more serious than selling our personal ators html/. information to advertisers.” Page 17 of 22 APP. 55 Exhibit E: Evidence of Implied Consent Source Bates Number Publication Summary 62 FB000000364 Emil Protalinski, Facebook Scans Chats The author of this article reports that “Facebook has added sleuthing to its array of data-mining capabilities, scanning and Posts for Criminal Activity, CNet (July your posts and chats for criminal activity.” He writes that Facebook’s “scanning program looks for certain phrases 12, 2012), available at found in previously obtained chat records from criminals, including sexual predators,” noting that the “relationship http://www.cnet.com/news/facebook-scans- analysis and phrase material have to add up before a Facebook employee actually looks at communications and makes chats-and-posts-for-criminal-activity/. the final decision of whether to ping the authorities.” The author also writes that “details of the tool are still scarce,” and that “Facebook likely wants to avoid discussing the existence of the monitoring technology in order to avoid further privacy concerns” because “[m]any users don’t like the idea of having their conversations reviewed, even if it’s done by software and rarely by Facebook employees.” 63 FB000000367 Alex Fitzpatrick, Facebook Monitors Your Chats for Criminal Activity, Mashable (July 12, 2012), available at http://mashable.com/2012/07/12/facebookscanning-chats/. This article notes that “Facebook and other social platforms are watching users’ chats for criminal activity.” The author writes that Facebook’s “screening process begins with scanning software that monitors chats for words or phrases that signal something might be amiss,” and that Facebook’s “scanning program is also ‘smart’ [because] it’s taught to keep an eye out for certain phrases found in the previously obtained chat records from criminals including sexual predators.” He suggests that “[k]eeping most of the scanned chats out of the eyes of Facebook employees may help Facebook deflect criticism from privacy advocates, but whether the scanned chats are deleted or stored permanently is yet unknown.” He ends by quoting Facebook’s then-Chief Security Officer Joe Sullivan, who says, “‘We’ve never wanted to set up an environment where we have employees looking at private communications, so it’s really important that we use technology that has a very low false-positive rate.’” 64 FB000000095 Emil Protalinski, Facebook is hiding your messages from you, ZDNet (December 10, 2011), available at http://www.zdnet.com/blog/facebook/faceb ook-is-hiding-your-messages-fromyou/6017/. This article concerns Facebook’s “Other” Messages folder and Facebook’s filtering mechanism for the Messages product. The author writes that the “Other” Messages folder is “supposed to work as a junk/spam folder, “ and that while he mainly received “mass” and “spam” messages in his own folder, other people were “really annoyed that Facebook classified some [messages] incorrectly because they missed important information.” The article quotes Facebook’s description of the new filtering service: “‘It seems wrong that an email from your best friend gets sandwiched between a bill and a bank statement . . . It’s not that those other messages aren’t important, but one of them is more meaningful. With new Messages, your Inbox will only contain messages from your friends and their friends. All other messages will go into an Other folder where you can look at them separately.’” The author writes, “I have no problem with such a folder existing: even my friends who say they missed an important message admit that most of their messages in there are not worth their time,” and he concludes by suggesting that Facebook make the “Other Messages” folder more “obvious.” 65 FB000000123 Elizabeth Weingarten, Furious At Facebook Again, Slate (December 9, 2011), available at http://www.slate.com/articles/technology/te chnology/2011/12/facebook_s_other_mess ages_mail_you_are_probably_missing html /. In this article, the author recounts a story where Facebook filtered a message into the “Other” Messages folder that she had wanted to receive. She indicates that she spoke with a Facebook representative, who explained Facebook’s filtering technology and that its move to a “Social Inbox” in November 2010 was a means for “sift[ing] out ‘meaningful’ messages from less meaningful ones.” The author quotes a Facebook statement about the “Other” Messages folder: “‘It seems wrong that an email message from your best friend gets sandwiched between a bill and a bank statement. It’s not that those other messages aren’t important, but one of them is more meaningful. With new Messages, your Inbox will only contain messages from your friends and their friends. All other messages will go into an Other folder where you can look at them separately.’” She concludes by asking, “So do I curse Facebook because it hid [the] messages, or praise it for allowing him to get in touch? I’m going to do both. Thanks, Facebook, for helping this nice man return my laptop. But please try to explain your services better. I suspect many people would be grateful ” Page 18 of 22 APP. 56 Exhibit E: Evidence of Implied Consent Source Bates Number Publication Summary 66 FB000000139 Dave Copeland, Facebook’s Email Reporting that “Facebook confirmed [that week] that it scans private messages for links and records them as likes, Scanning Isn’t A Privacy Issue, It’s A according to the Wall Street Journal and other news outlets,” the author opines that this practice injures Facebook’s Credibility Issue, Read Write (October 5, credibility. The author writes that “Facebook has not kept secret its scanning of private messages for references to 2012), available at criminal activity,” but he suggests that it is “new [ ] that it also looks for links and records those at likes,” which he http://readwrite.com/2012/10/05/facebooks- claims “gives the appearance that more people are liking more things on the social network.” Quoting Facebook’s email-scanning-isnt-a-privacy-issue-its-a- statement that the “scanned links were counted as engagement, not endorsement,” the author opines that Facebook’s credibility-issue/. statement “misses the point” because “Facebook’s practice of scanning messages and counting links as likes isn’t a privacy issue.” He writes that “[i]t’s common knowledge that what users do online—even in so-called private messaging—is potentially public.” Instead, the author opines, “Facebook’s activity raises a credibility issue” because “[i]t shows that the company is fudging the numbers when it comes to advertising.” 67 FB000000209 Jim Edwards, This Flaw In Facebook Lets You Create As Many Fake Likes As You Want, Business Insider (October 5, 2012), available at http://www.businessinsider.com/this-flawin-facebook-lets-you-create-as-many-fakelikes-as-you-want-2012-10/. Here, the author writes that his publication previously reported that Facebook “quietly scans your messages, searching for URLs that you’ve sent to your friends [and] [w]hen it sees one, it increases the number of Facebook Likes on that URL.” He notes that while other publications have “portrayed this as a privacy invastion,” he believes that “more importantly, it appears to be a massive source of bogus Likes.” The author goes on to describe the Facebook bug that increased the social plugin count on third-party websites by two instead of one, and he includes an example of an instance where he attempted to “generate some fake clicks.” He quotes a Facebook spokesperson’s statement acknowledging that the company “did recently find a bug with our social plugins where at times the count for the Share or Like goes up by two” and that the company was working on a “fix” for that issue, and the author notes that “those likes may actually reflect negative consumer sentiment.” 68 FB000000226 Social Networks: Can Robots Violate User Privacy?, High-Tech Bridge (August 7, 2013), available at https://www htbridge.com/news/social_net works_can_robots_violate_user_privacy ht ml/. This article indicates that High-Tech Bridge conducted a “a simple technical experiment to verify how the 50 largest social networks, web services and free emails systems respect—or indeed abuse—the privacy of their users.” The article purportedly confirms that Facebook, Twitter, Google+, and Formspring are crawling URLs, with Facebook allegedly crawling a “[p]rivate message with a link.” The article “tak[es] into consideration that some of the services may have legitimate robots (e.g. to verify and block spam links) crawling every user-transmitted link automatically,” and it notes that High-Tech Bridge “created a robots.txt file on our web server that restricted bots accessing the server and its content.” There are several comments to the article written by readers, one of which notes, “‘[N]ot surprised at all Facebook did it. Just take note that Facebook do[es] scrap[e] URLs . . . to construct previews . . . [and] to run those URLs against a malicious signaled table of URLs for the ‘protection’ of their production and users. I can be wrong but I see no space here for a lawsuit.’” 69 FB000000289 Facebook ‘Likes’ Automatically Added Without User-Clicks, BBC News Technology (October 4, 2012), available at http://www.bbc.com/news/technology19832043/. This article reports that a security researcher recently “found that simply sending a web address to a friend using Facebook’s private messaging function would add two likes to that page.” The article quotes portions of a statement from a Facebook spokesperson confirming that Facebook discovered a bug where the Share or Like goes up by two and that Facebook was “‘working on a fix to solve the issue’” and that “‘no user information is exchanged.’” The article continues by noting that “[i]n documentation relating to the function of the like button, Facebook details four criteria which cause the likes number to increase - only one of which involves clicking the like button.” It further reads, “Facebook stressed that the added likes were anonymous, and would not appear on the user’s timeline.” Page 19 of 22 APP. 57 Exhibit E: Evidence of Implied Consent Source Bates Number Publication Summary 70 FB000000296 Craig Lloyd, Facebook Auto-Liking Pages for Users Without Permission [Updated], Slash Gear (October 4, 2012), available at http://www.slashgear.com/facebook-autoliking-pages-for-users-without-permission04250415/. In this article, the author cites reports indicating that “Facebook is scanning its users’ private messages and searching for links to Facebook fan pages . . . [and] supposedly automatically likes the pages for you without asking for your permission to do so.” The author writes that “it only seems that it increases the Like count of a page, and doesn’t actually ‘like’ the page on your behalf,” noting that “some users are reporting that it actually does like the page for you without your permission.” The author opines that “this can be a huge problem,” but he indicates that “[s]canning itself is nothing new” and notes that “Gmail does it to provide its users with targeted ads.” The author concludes by noting that “this auto-liking debacle takes it to another level that’s a little over the line and unnecessary.” He updates the article to include a Facebook spokesperson’s statement on the issue, which indicates, among other things, that “‘[a]bsolutely no private information has been exposed and Facebook is not automatically Liking any Facebook Pages on a user’s behalf” and that “[w]hen the count is increased via shares over private messages, no user information is exchanged, and privacy settings of content are unaffected.’” 71 FB000000298 Ryan Singel, Juking Your Facebook ‘Like’ Stats Is As Easy As Sending a Message, Wired (October 4, 2012), available at http://www.wired.com/2012/10/facebooklikes-messages/. The author of this article reports that people “looking to artificially inflate their Facebook stats . . . can just simply send a raft of private messages that include a link to your page, and Facebook will add +2 to your page’s ‘Like’ count for each message.” He acknowledges that “[i]t’s long been known that Facebook scans internal messages for spam and security risks—and that it blocks users from sending links to torrent sites such as The Pirate Bay,” but that it’s “never been clear how much data mining [Facebook] is doing” of users’ Facebook messages. He summarizes Ashkan Soltani’s Wall Street Journal article on the topic, and then writes that Facebook’s behavior is “not a bug,” but is “something actually noted in the documentation for developers.” He updates his article to include Facebook’s spokesperson’s statement on the issue, which reads as follows: “‘Absolutely no private information has been exposed and Facebook is not automatically Liking any Facebook Pages on a user’s behalf. Many websites that use Facebook’s “Like”, “Recommend”, or “Share” buttons also carry a counter next to them. This counter reflects the number of times people have clicked those buttons and also the number of times people have shared that page’s link on Facebook. When the count is increased via shares over private messages, no user information is exchanged, and privacy settings of content are unaffected. Links shared through messages do not affect the Like count on Facebook Pages.’” 72 FB000000340 Ed Oswald, Facebook Private Messages Trigger ‘Likes’ Without Telling, TechHive (October 5, 2012), available at http://www.techhive.com/article/2011278/f acebook-private-messages-trigger-likeswithout-telling html/. In this article, the author notes that “[t]he next time you share a link with a Facebook friend via private message, be aware that you’re anonymously ‘liking’ that page publicly as well.” He writes that “[w]hile this may come as a surprise, evidence that the company was scanning our messages for these likable links has been public for at least a week,” citing Facebook’s “September 27 FAQ for developers that [states that] ‘the number of inbox messages containing this URL as an attachment’ is a factor in counting the number of likes that shows up on a page’s Like Button,’” among other factors. The author writes that “[w]hile this information seems to have been public for some time, those of us who aren’t developers likely had no clue of Facebook’s actions,” but that “given how Facebook uses our activities to further its own business interests, this practice shouldn’t surprise us.” Page 20 of 22 APP. 58 Exhibit E: Evidence of Implied Consent Source Bates Number Publication Summary 73 FB000000342 Lisa Vaas, Facebook Scans Private Messages to Inflate the ‘Like’ Counter on Websites, Naked Security (October 8, 2012), available at https://nakedsecurity.sophos.com/2012/10/ 08/facebook-scans-private-messages-likecounter/. Here, the author writes that “Facebook has confirmed that it’s scanning private Facebook messages to boost ‘Like’ counters on third party websites.” She continues by noting that Facebook “confirmed that they had discovered a bug affecting Like counts” that “concerned inflating page counts by two Likes instead of one.” She reports that “[t]he fact that this is [sic] function is baked into Facebook code as opposed to being a potential fluke of privacy transgression is confirmed . . . on the Facebook Developers page, which states that a websites’ number of Likes is the sum of: [t]he number of likes of this URL; [t]he number of shares of this URL (this includes copy/pasting a link back to Facebook); [t]he number of likes and comments on stories on Facebook about this URL; [t]he number of inbox messages containing this URL as an attachment.” The author confirms that “Facebook’s scanning of private messages isn’t new” and that “[t]he power of the social media mammoth’s data mining technology when applied to private messages came to light in March, when Facebook was credited with quashing potential child molestation between a 13-year-old girl and a man in his 30s who were having a private Facebook conversation about sex.” She writes that Facebook’s “data mining technology scans postings and chats for criminal activity, analyzing relationships to find suspicious conversations between unlikely pairings.” Noting that “[e]mail providers such as Gmail also have a long-standing practice of reviewing messages to weed out spam and to target ads,” she opines that “[t]hose are reasonable uses of data mining technology, but it’s disconcerting to find what might be yet more intrusive forays into allegedly private messages.” For that reason, she writes that “it’s a bit of a relief to learn that Facebook later clarified the privacy issue, saying that ‘absolutely no private information’ is exposed in the private-message-derived Like inflation.” She concludes by writing, “Be prepared to add to your subjects’ Facebook counter glow, whether you want to or not, if you send URLs via private Facebook conversations.” 74 FB000000353 Joseph Menn, Social Networks Scan for Sexual Predators, With Uneven Results, Reuters (July 12, 2012), available at http://www reuters.com/article/2012/07/12/ us-usa-internet-predatorsidUSBRE86B05G20120712/. This article discusses “Facebook’s extensive but little-discussed technology for scanning postings and chats for criminal activity,” which flagged a sexual predator chatting with a 13 year-old girl. The author writes that “Facebook is among the many companies that are embracing a combination of new technologies and human monitoring to thwart sex predators,” noting that “[s]uch efforts generally start with automated screening for inappropriate language and exchanges of personal information.” He writes that “[l]ike most of its peers, Facebook generally avoids discussing its safety practices to discourage scare stories, because it doesn’t catch many wrongdoers, and to sidestep privacy concerns [because] [u]sers could be unnerved about the extent to which their conversations are reviewed, at least by computer programs.” The author continues by noting that “[i]n part because of its massive size, Facebook relies more than some rivals on such technology.” He quotes Facebook’s then-Chief Security Officer Joe Sullivan, who says, “‘We’ve never wanted to set up an environment where we have employees looking at private communications, so it’s really important that we use technology that has a very low false-positive rate.’” 75 FB000000402 Chi Ibe, Nowhere to Hide: Facebook Monitors Your Chats, YNaija.com (July 13, 2012), available at http://ynaija.com/the-internet-has-noprivacy-settings-facebook-monitors-yourchats/. Writing that “[r]eports have revealed that Facebook and other social platforms are watching users’ chats” to “monitor criminal activity and notifying police if any suspicious behaviour is detected,” the author asks, “[W]hat ever [sic] happened to good old privacy?” She indicates that “a number of social networking sites have set up a screening process which works by a scanning software that monitors chats for words or phrases that signal something might be amiss.” Page 21 of 22 APP. 59 Exhibit E: Evidence of Implied Consent Source Bates Number Publication Summary 76 FB000000407 Tim Bukher, Facebook Monitoring User The author writes that “[a]ccording to a report via Mashable, Facebook does more than passively scan user profile Chats, Reporting to Police, LawTechie settings for targeted advertising, it also monitors chats between users for potential criminal activity.” (July 13, 2012), available at http://www.lawtechie.com/2012/07/facebo ok-monitoring-user-chats-reporting-topolice/. 77 FB000000409 Fox 13 Tampa Bay Staff, Facebook Uses Technology to Spy on Private Chats, My Fox - Tampa Bay (July 13, 2012), available at http://www myfoxtampabay.com/story/190 17765/2012/07/13/facebook-usestechnology-to-spy-on-private-chats/. This article reports that “Facebook’s chief security officer admits Facebook users are being monitored for any suspected criminal activity, and it’s not just the stuff you post on timelines.” It notes Facebook uses “smart software” to “monitor[ ]personal chats” and “scans those chats for certain phrases, exchanges of personal information and vulgar language.” The article adds that “Facebook says the technology has a very low false-positive rate to protect its users’ privacy, but as expected there has been a backlash from users” who “feel their private conversations are being violated.” The article acknowledges that Facebook’s technology “helped net an alleged sexual predator” and reports that the FBI is “on board with this technology and hopes more online sites use it.” Page 22 of 22 APP. 60 EXHIBIT F APP. 61 APP. 62 APP. 63 APP. 64 APP. 65 APP. 66 APP. 67 APP. 68 APP. 69 APP. 70 APP. 71 APP. 72 APP. 73 APP. 74 APP. 75 APP. 76 APP. 77 APP. 78 APP. 79 APP. 80 APP. 81 APP. 82 APP. 83 APP. 84 APP. 85 APP. 86 APP. 87 APP. 88 APP. 89 APP. 90 APP. 91

Disclaimer: Justia Dockets & Filings provides public litigation records from the federal appellate and district courts. These filings and docket sheets should not be considered findings of fact or liability, nor do they necessarily reflect the view of Justia.


Why Is My Information Online?