Kennedy Jr. v. Meta Platforms, Inc.

Opinion

ROBERT F. KENNEDY, et al., Case No. 3:24-cv-02869-WHO

Plaintiffs, ORDER DENYING MOTION FOR v. PRELIMINARY INJUNCTION

META PLATFORMS, INC., et al., Re: Dkt. Nos. 29, 40 Defendants.

Former presidential candidate Robert F. Kennedy Jr., a political action committee supporting his campaign called American Values 2024 (“AV24”), and individual Jessica Reed Kraus (collectively, the “plaintiffs”) filed this action against Meta Platforms, Inc., Facebook Operations, LLC, Instagram, LCC, and Mark Zuckerberg (collectively, the “defendants”), alleging violations of their rights under the First Amendment, the Voting Rights Act (“VRA”), and the Ku Klux Klan Act. In brief, the plaintiffs allege that the defendants have conspired with the federal government to remove and block from the defendants’ social media platforms the content and users that express support for Kennedy’s campaign. The plaintiffs move for a preliminary injunction, requesting “immediate” relief to stop the alleged censorship. Because the plaintiffs fail to establish a likelihood of success on the merits of any of their claims, and for the following reasons, the motion is denied. Kennedy is a presidential candidate during the 2024 election cycle, ending his campaign shortly before this motion was heard but remaining on the ballot in non-swing states. See Amended Complaint (“AC”) [Dkt. No. 28] ¶¶ 2, 87. AC24 is a political action committee “invested significant time and resources following both former President Trump and Mr. Kennedy on the campaign trail.” Id. ¶¶ 70, 72, 89. AC24 has an account on Facebook, and Reed Kraus has an account on Instagram. See id. ¶¶ 28, 72. Facebook and Instagram, two of the defendants here, are owned by defendant Meta, and Zuckerberg is Meta’s Chief Executive Officer. Id. ¶¶ 90–93. In May 2024, AC24 produced and released a thirty-minute “documentary film” about Kennedy called “Who Is Bobby Kennedy?” that “invit[ed] voters to make up their own minds about Mr. Kennedy, rather than accepting falsehoods about him repeatedly asserted by major news outlets and social media platforms.” AC ¶¶ 23–24. The plaintiffs allege that they posted links to the video on Facebook and Instagram, which were blocked or removed within ten minutes so that users could watch, share, or post links to the video. Id. ¶¶ 31–35. The plaintiffs allege that users who tried to share the video received a variety of messages from the defendants about why the video was banned, including for violating community standards, being spam, promoting crime or hate groups, soliciting sexual services and firearms, being malicious, or containing banned COVID-19 misinformation content. Id. ¶¶ 37, 40. The plaintiffs deny that any of these reasons are an accurate portrayal of the video. Id. ¶¶ 38–39, 41, 51. They say that this censorship prevented millions of people from seeing the film, like the one hundred million views the video received on X (formerly Twitter). Id. ¶¶ 42, 72–73. The plaintiffs allege that the suppression of this video has caused and is causing: “substantial donation losses” to Kennedy and AV24; “substantial injury” to Kennedy’s candidacy, to his and AV24’s free speech rights, and to AV24’s property rights; and “substantial injury to the rights of citizens and voters who sought to express their support” to Kennedy. Id. ¶ 57. The defendants submitted declarations of Meta employees who work in content moderation and enforcement divisions of the company. See Declaration of Sonal Mehta (“Mehta Decl.”) [Dkt. No. 35-1] Exs. A, B. One employee explained that Meta did prevent the video from being posted on May 4 because of a “mistaken[] determin[ation] that it violated Meta’s policies.” Id. Ex. A ¶ 6. The employee said this happens when the automated content moderation system The employee explained that this can happen with new links that are posted to the platforms for the first time or when the poster tags multiple users in posts containing the link. Id. Access was restored within thirty minutes of Meta being alerted to the problem. Id. ¶ 8. The post remains available on AV24’s Facebook page. Mehta Decl. Ex. N. The plaintiffs also assert that Reed Kraus’ content about Kennedy has “been widely censored and demoted,” though her content about Donald Trump has not. AC ¶ 72. When she wrote a pro-Kennedy post in October 2023, she lost 40,000 followers overnight and says “[t]here is no explanation for this unprecedent phenomenon other than” the defendants’ “surreptitious ‘demoting’ or ‘de-boosting’ of her Instagram account.” Id. ¶¶ 73–74. The following month, she says fewer people viewed her Instagram story, which she attributes to the defendants’ “shadow- banning” of her content. Id. ¶ 75. After posting the Who Is Bobby Kennedy video, she was locked out of her account and the account was precluded from being shown to new followers. Id. ¶ 80. Instagram allegedly also blocked a poll she posted about Kennedy debating as violative of community standards. Id. ¶ 84. The defendants’ declaration states that “no enforcement actions” were taken against any of Reed Kraus’ posts about Kennedy or “RFK.” Mehta Decl. Ex. B ¶ 5. The Meta employee says that Reed Kraus’ suspension was due to violation of Meta’s policies about physical harm. Id. ¶ 7. A post was removed and Reed Kraus was suspended for less than 13 minutes because she wrote, “[I]f I show up on campus and see my kid in a facemask spray painting the school I broke my back to send him to, demanding vegan food and denouncing bagels, I’m spanking him in front of the whole student body, telling him to ‘stop being an idiot.’” Id. ¶¶ 6–10. Finally, the plaintiffs provide screenshots of third parties that apparently show censorship of pro-Kennedy content on their Facebook and Instagram pages. See [Dkt. No. 29-1] Exs. A, B; AC ¶¶ 61–64 (making general assertions that third parties’ posts and content were censored, without identifying the third parties, posts, or content). In response, the defendants submitted an administrative motion for leave to file supplemental evidence, asserting that looking into every screenshot was time- and resource-intensive, and that they were only recently able to confirm the The plaintiffs filed a motion for a preliminary injunction, asserting they are likely to succeed on the merits of their First Amendment, Voting Rights Act, and Ku Klux Klan Act claims. (“Mot.”) [Dkt. No. 29]. The defendants opposed. (“Oppo.”) [Dkt. No. 35]. The plaintiffs replied. (“Repl.”) [Dkt. No. 39]. I held a hearing at which counsel for both parties appeared. “[A] party is entitled to a preliminary injunction if it demonstrates (1) ‘that [it] is likely to succeed on the merits,’ (2) ‘that [it] is likely to suffer irreparable harm in the absence of preliminary relief,’ (3) ‘that the balance of equities tips in [its] favor,’ and (4) ‘that an injunction is in the public interest.’” Flathead-Lolo-Bitterroot Citizen Task Force v. Montana, 98 F.4th 1180, 1190 (9th Cir. 2024) (quoting All. for the Wild Rockies v. Cottrell, 632 F.3d 1127, 1131 (9th Cir. 2011)); see also Winter v. Nat. Res. Def. Council, Inc., 555 U.S. 7, 20 (2008). The Ninth Circuit also uses a “‘sliding scale’ variant of the Winter test.” Flathead-Lolo-Bitterroot, 98 F.4th at 1190. That scale allows the plaintiff to show entitlement to a preliminary injunction by making a “lesser showing” of “serious questions going to the merits,” so long as the plaintiff shows that the “balance of hardships . . . tips sharply towards the plaintiff,” and the other two factors are met. Id. (emphasis added) (citations omitted). I. LIKELIHOOD OF SUCCESS ON THE MERITS OR SERIOUS QUESTIONS A. First Amendment Claim The plaintiffs assert that they are likely to succeed on the merits of their First Amendment claim, which is that Meta violated their rights to free speech by censoring their posts and accounts on Meta’s platforms. But the First Amendment “‘prohibits only governmental abridgment of speech’ and ‘does not prohibit private abridgment of speech.’” Children’s Health Def. v. Meta Platforms, Inc., ---F. 4th---, No. 21-16210, 2024 WL 3734422, at *4 (9th Cir. Aug. 9, 2024) (first quoting Manhattan Cmty. Access Corp. v. Halleck, 587 U.S. 802, 808 (2019); and then citing Prager Univ. v. Google LLC, 951 F.3d 991, 996 (9th Cir. 2020)). Because there is no apparent “To begin by stating the obvious, Meta, the owner of Facebook [and Instagram], is a private corporation, not a government agency.” Id. “[I]n certain ‘exceptional cases,’ a private party ‘will be treated as a state actor for constitutional purposes.’” Id. (quoting O’Handley v. Weber, 62 F.4th 1145, 1155–56 (9th Cir. 2023)). “The private party must meet two distinct requirements: (1) the ‘state policy’ requirement and (2) the ‘state actor’ requirement.’” Id. (first quoting Wright v. Service Emps. Int’l Union Loc. 503, 48 F.4th 1112, 1121 (9th Cir. 2022); then citing Lugar v. Edmondson Oil Co., 457 U.S. 922, 937 (1982); and then citing O’Handley, 62 F.4th at 1156). The plaintiffs argue that the defendants meet both requirements. I address them in turn. 1. State Policy “To satisfy the state policy requirement, the alleged constitutional deprivation must result from ‘the exercise of some right or privilege created by the State’ or ‘a rule of conduct imposed by the State or by a person for whom the State is responsible.’” Children’s Health, 2024 WL 3734422, at *4 (quoting Lugar, 457 U.S. at 937). First, courts “look to whether the source of the alleged constitutional harm is a state statute or policy.” Id. (quoting Belgau v. Inslee, 975 F.3d 940, 947 (9th Cir. 2020)) (quotation marks omitted). “This requirement is satisfied when a private institution ‘enforce[s] a state-imposed rule’ instead of ‘the terms of its own rules.’” Id. (quoting O’Handley, 62 F.4th at 1156). The analysis begins with “identifying the ‘specific conduct of which the plaintiff complains,’” which requires determining the “source” of any alleged harm. Id. at *6 (first quoting Wright, 48 F.4th at 1122; and then quoting Ohno v. Yasuma, 723 F.3d 984, 994 (9th Cir. 2013)). The Ninth Circuit recently has twice affirmed dismissal of claims filed by plaintiffs alleging that social media platforms violated the plaintiffs’ First Amendment rights by flagging, removing, or otherwise “censoring” the plaintiffs’ content shared on those platforms. See Children’s Health, 2024 WL 3734422 at *2–4; O’Handley, 62 F.4th at 1153–55. In both cases, the Ninth Circuit held that the plaintiffs’ claims failed at the first step of the state action framework because of “the simple fact” that the defendants “acted in accordance with [their] own see also Children’s Health, 2024 WL 3734422 at *5–6. The only difference between those cases and this one is that here, the plaintiffs seem to allege that the “specific” harmful conduct is Meta’s censorship itself, rather than its policy of censoring. Based on the documents submitted and allegations made, that is a distinction without a difference. Meta shows that it removed plaintiff Reed Kraus’ posts and suspended her account for less than 13 minutes because she violated its terms and conditions related to her threats of physical violence. Mehta Decl. Ex. B ¶¶ 6–10. Meta’s right to take these actions “derive[s]”—as in O’Handley and Children’s Health—“from its user agreement[s],” not from any state-imposed rule. O’Handley, 62 F.4th at 1126. To the extent that the plaintiffs allege Meta censored pro-Kennedy content posted by nonparties, see Mot. 1:9–2:6, AC ¶ 8, [Dkt. No. 29-1] Exs. A, B—and setting aside the standing issues unaddressed by the plaintiffs about their ability to assert that their First Amendment rights were violated by Meta’s censorship of others’ accounts—they fail to plead facts or provide evidence showing that these actions were taken in accordance with a state- imposed rule. At bottom, the plaintiffs fail to point to any governmental “statute or policy” as the source of their harm. Children’s Health, 2024 WL 3734422 at *4. To support their theory, they rely on two district court orders from Louisiana and an “Interim Staff Report” for House of Representatives’ Committee on the Judiciary.1 See Mot. 21:8–16; Repl. 11:7–13:7; AC ¶¶ 158– 62; Missouri v. Biden, 680 F. Supp. 3d 630 (W.D. La.), aff’d in part, rev’d in part, 80 F.4th 641 (5th Cir. 2023), opinion withdrawn and superseded on reh’g, 83 F.4th 350 (5th Cir. 2023), rev’d and remanded sub nom. Murthy v. Missouri, 144 S. Ct. 1972 (2024); Kennedy v. Biden, No. 3:23- CV-00381, 2024 WL 625327 at *10 (W.D. La. Feb. 14, 2024) (appeal pending). These orders and interim report discussed actions taken by social media companies and various federal government officials and agencies in 2020 through 2021 with respect to removing COVID-19 misinformation

1Comm. on the Judiciary, U.S. H.R. Interim Staff Rep., The Censorship-Industrial Complex: How Top Biden White House Officials Coerced Big Tech to Censor Americans, True Information, and Critics of the Biden Administration (May 1, 2024), https://judiciary.house.gov/sites/evo- from social media platforms. See Missouri, 680 F. Supp. 3d at 642–89; Kennedy, 2024 WL 625327, at *10 (“Kennedy Plaintiffs rely on the same evidence presented in Missouri v. Biden.”). Missouri was ultimately reversed by the Supreme Court, which held that the plaintiffs lacked standing to pursue their claims. Murthy, 144 S. Ct. at 1981. In Murthy, the Supreme Court analyzed Missouri, where the plaintiffs were “two states and five social-media users” who sued “dozens” of federal government officials and agencies for allegedly pressuring social media platforms to suppress speech protected by the First Amendment. Id. The allegedly censored speech was flagged and removed as misinformation concerning the COVID-19 pandemic. See id. at 1982–85. The Court held that the plaintiffs failed to establish standing because the evidence showed that the social media platforms made and enforced the content moderations decisions, and no findings of fact were made showing “specific causation” between the government actions and the content moderation. Id. at 1987. The Fifth Circuit had determined that the alleged censorship was “likely attributable at least in part” to the actions of the government defendants, which the Supreme Court held was an “overly broad assertion” because the evidence actually showed that the defendants both made similar decisions before any government actors were involved and also declined to follow the government’s alleged requests for censorship on multiple occasions. Id. The Supreme Court added that the Fifth Circuit “glossed over complexities in the evidence” in part because it “relied on the District Court’s factual findings, many of which unfortunately appear to be clearly erroneous.” Id. at 1987 & n.4. The Court then reviewed and cited the record to show how the district court’s findings were incorrect. Id. at 1987 n.4. The Murthy opinion makes my decision here straightforward. Murthy rejected Missouri’s factual findings and specifically explained that the Missouri evidence did not show that the federal government caused the content moderation decisions. Yet here, the plaintiffs rely on Missouri as their evidence that a state rule caused the defendants’ alleged censorship actions. Even if I accepted the vacated district court order as evidence here—which I do not—the Supreme Court has plainly explained why it does not support the plaintiffs’ argument. government had a policy in 2020 and 2021 related to COVID-19 misinformation. They say nothing about whether any government had a policy in 2024 related to Kennedy’s campaign. The plaintiffs are therefore without evidence of any government policy, which means they cannot show their harm is derived from such a policy. See Children’s Health, 2024 WL 3734422 at *4; see also Kennedy v. Google LLC, 688 F. Supp. 3d 951, 958 (N.D. Cal. 2023), affirmed, No. 23-3411, 2024 WL 3934326, at *1 (9th Cir. Aug. 26, 2024) (declining to use Missouri factual findings as evidence, in a similar case brought by plaintiff Kennedy against another social media company for content moderation decisions). Accordingly, as in Children’s Health and O’Handley, the plaintiffs’ First Amendment claim “fails at this threshold step.” Children’s Health, 2024 WL 3734422 at *6. And because this is “fatal” to the plaintiffs’ state action claim, id. (citing Lindke v. Freed, 601 U.S. 187, 198 (2024)), the plaintiffs fail to show likelihood of success on the merits or serious questions going to the merits of their First Amendment claims, see Flathead-Lolo-Bitterroot, 98 F.4th at 1190. 2. State Actor Though I need not address the second prong of the test, see Children’s Health, 2024 WL 3734422 at *6, I follow the direction of the Ninth Circuit in Children’s Health and explain why the plaintiffs’ claims fail to satisfy the second part of the state action test as well. “[T]he Supreme Court has identified four tests for when a private party ‘may fairly be said to be a state actor’: (1) the public function test, (2) the joint action test, (3) the state compulsion test, and (4) the nexus test.” Id. (citing Lugar, 457 U.S. at 937, 939). Here, the AC does not clarify which test or tests that the plaintiffs rely upon, and the plaintiffs’ briefing fails to discuss this second prong at all. The plaintiffs clearly do not argue that Meta “performs a traditionally public function.” Id. at *5 (quoting Manhattan Cmty. Access, 587 U.S. at 804). They do seem to allege that Meta “is a ‘willful participant in joint activity’ with the government” under the joint action test. Children’s Health, 2024 WL 3734422, at *5 (quoting Lugar, 457 U.S. at 941); see AC ¶ 164 (citing Justice Alito’s dissent in Murthy, in which he said that “the platforms have complied” with the government’s requests), ¶ 165 (describing how social 177, 181, 184 (describing “collusion” and “work[ing] jointly”). And the plaintiffs also seem to allege the state compulsion test, that “the government compels or encourages [Meta] to take a particular action.” Children’s Health, 2024 WL 3734422, at *5 (citing Blum v. Yaretsky, 457 U.S. 991, 1004 (1982)); see AC ¶ 158, 179–80, 184 (describing the Louisiana district court’s findings that the government’s actions constituted “coercion, significant encouragement, and collusion” and noting “federal pressure”). It also seems possible that the plaintiffs allege that “there is a ‘sufficiently close nexus’ between the government and the challenged action” for the nexus text. Children’s Health, 2024 WL 3734422, at *5 (quoting Jackson v. Metropolitan Edison Co., 419 U.S. 345, 351 (1974)). The plaintiffs’ theory is that Meta and the government colluded or acted jointly, or the government coerced Meta, to remove content related to Kennedy’s 2024 presidential campaign from Meta’s platforms. The problem with that theory is again the lack of evidence. The Missouri and Kennedy findings were rejected by the Supreme Court, as explained above. And they—and the interim report—suggest at most a relationship or communications between Meta and the government about removal of COVID-19 misinformation in 2020 and 2021. Even if the plaintiffs proved that Meta and the government acted jointly, or colluded, or that Meta was coerced by the government to remove and flag COVID-19 misinformation three years ago, that says nothing about Meta’s relationship and communications with the government in 2024. Nor does it suggest that Meta and the government worked together to remove pro-Kennedy content from Meta’s platforms. Because of this, the plaintiffs fail to show likelihood of success on the merits—or serious questions going to the merits—for any of the three possible state action prongs. They do not provide evidence or allegations of a “specific[]” agreement between Meta and the government to specifically accomplish the goal of removing Kennedy content from Meta platforms. See Children’s Health, 2024 WL 3734422, at *5 (describing joint action test and collecting cases). Nor do they show that the government exercised coercive power or “significant encouragement” for Meta to remove Kennedy-related content in 2024. Id. at *9–10 (describing coercion test and on social media sites was insufficient to show government coerced platforms to remove it). And for similar reasons, the plaintiffs do not establish a “sufficiently close nexus” between the government and the removal of Kennedy-related content from Meta’s platforms. Id. at *5. Their First Amendment claim accordingly fails at step two of the state action inquiry. It is far from likely to succeed on the merits. B. Statutory Claims The plaintiffs assert that they are likely to succeed on the merits of two of their statutory claims, for violation of §11(b) Voting Rights Act (“VRA”), 52 U.S.C. § 10307(b), and for violation of the Support or Advocacy Clause of the Ku Klux Klan Act of 1870, now codified at 42 U.S.C. § 1985(3). I address these claims in turn, and then briefly discuss the defendants’ argument that they are immunized under § 230 of the Communications Decency Act. 1. Voting Rights Act In relevant part, § 11(b) of the VRA provides, “No person, whether acting under color of law or otherwise, shall . . . intimidate, threaten, or coerce . . . any person for urging or aiding any person to vote . . . .” 52 U.S.C.A. § 10307(b). The statute is broad and protects both the right to vote and the right to urge others to vote. See id.; Allen v. City of Graham, No. 1:20-CV-997, 2021 WL 2223772, at *8 (M.D.N.C. June 2, 2021). It is well established that § 11(b) “reaches beyond government actors, affording a private right of action,” and reaching private conduct. Nat’l Coal. on Black Civic Participation v. Wohl, 661 F. Supp. 3d 78, 112 (S.D.N.Y. 2023) (“Wohl III”) (collecting cases). A district court in New York recently “examined at length the conduct encompassed by the terms ‘intimidate,’ ‘threaten,’ and ‘coerce’ under the statute.” Id. at 113 (explaining, in its order on summary judgment, the court’s prior analyses in orders addressing motions for a temporary restraining order and to dismiss). I am persuaded by that court’s extensive, thorough analysis and agree with its definition of these terms. See also Fair Fight Inc. v. True the Vote, No. 2:20-CV- 00302-SCJ, 2024 WL 24524, at *37–38 (N.D. Ga. Jan. 2, 2024) (citing and accepting this definition from Wohl II); Ariz. All. for Retired Americans v. Clean Elections USA, 638 F. Supp. 3d (9th Cir. Jan. 26, 2023) (same, in a now-vacated order). “[I]ntimidation includes messages that a reasonable recipient, familiar with the context of the communication, would view as a threat of injury to deter individuals from exercising their right to vote.” Wohl III, 661 F. Supp. 3d at 113. “[U]nlawful threats or intimidation . . . need not be violent or physical, and may include communications inspiring fear of legal consequences, economic harm, dissemination of personal information, and surveillance.”2 Id. (first citing Nat’l Coal. on Black Civic Participation v. Wohl, 498 F. Supp. 3d 457, 477 (S.D.N.Y. 2020) (“Wohl I”); and then citing Nat’l Coalition on Black Civic Participation v. Wohl, 512 F. Supp. 3d 500, 509 (S.D.N.Y. 2021) (“Wohl II”)). Examples of threatened “legal consequences” that a reasonable person, familiar with context, would view as a threat of injury to deter her from voting include: threats that police will use voter information to find and enforce old warrants3; a pattern of baseless arrests at a voter registration event4; threats of deportation against lawful citizens who are immigrants and their family members5; and threats of filing a lawsuit against someone or suspending them without pay.6 See Wohl III, 661 F. Supp. 3d at 113–14. Examples of threatened “economic harm” that a reasonable person, familiar with context, would view as a threat of injury to deter her from voting include threats of “adverse economic consequences,” such as: threatening to personal information to debt collectors to collect outstanding debt7; evicting or threatening to evict sharecroppers from their land or otherwise 2 Though of course, “[t]he use of physical violence” such as beatings or pepper spray “to deter an individual from voting or engaging in a voting-related activity” also gives rise to a claim under § 11(b). Allen, 2021 WL 2223772, at *7 (citing Katzenbach v. Original Knights of Ku Klux Klan, 250 F. Supp. 330, 341 (E.D. La. 1965)).

3 Wohl III, 661 F. Supp. 3d at 113–14.

4 United States v. McLeod, 385 F.2d 734, 740–41 (5th Cir. 1967).

5 United States v. Tan Duc Nguyen, 673 F.3d 1259, 1265 (9th Cir. 2012) (analyzing threats under California election law).

6 Lovejoy-Wilson v. NOCO Motor Fuel, Inc., 263 F.3d 208, 222–23 (2d Cir. 2001) (addressing threats and intimidation in the context of the Americans with Disabilities Act). interfering with their sharecropping contracts8; inducing local merchants to boycott and refuse to provide necessities to anyone who helps register voters9; and prohibiting an insurance broker from accessing land to talk to his clients, and therefore preventing him from carrying out his profession, because of the broker’s voter registration efforts.10 Wohl III, 661 F. Supp. 3d at 114. Additionally, publishing names, addresses, and phone numbers of registered voters in a report accusing them of committing various felonies also constitutes threats and intimidation under § 11(b), both because of the “clear effort to subject the named individuals to public opprobrium” and the “fear of harassment and interference with their right to vote” that the named individuals experienced. League of United Latin Am. Citizens - Richmond Region Council 4614 v. Pub. Int. Legal Found., No. 1:18-CV-00423, 2018 WL 3848404, at *4 (E.D. Va. Aug. 13, 2018) (“LULAC”).11 Though I do not doubt that it is theoretically possible for someone to show they were intimidated, threatened, or coerced by a social media site for urging people not to vote, in violation of § 11(b), and in a way that parallels any of the above cases, the plaintiffs have not shown that is what happened here. First, with respect to blocking links to the Who Is Bobby Kennedy video, the parties

8 United States v. Beaty, 288 F.2d 653, 656–58 (6th Cir. 1961) (per curiam).

9 Beaty, 288 F.2d at 654; see also Katzenbach, 250 F. Supp. at 348 (citing “United States v. Deal, W.D. La. 1961, 6 Race Rel. L. Rep. 474”).

10 United States v. Bruce, 353 F.2d 474, 476–77 (5th Cir. 1965).

11 The plaintiffs argue that “disseminating ‘false utterances’ intended to chill, or having the effect of chilling, individuals from engaging in the protected activity” constitutes intimidation. Mot. 14:8–16. What they cite, though, is a pretrial decision about a criminal charge for 18 U.S.C. § 241, where the court held “that prosecutions of § 241 for ‘conspiracies to make verifiably false utterances about the time, place, or manner of elections that would injure the right to vote [are] unlikely to encourage selective prosecutions or chill broad categories of constitutional speech,’” a decision that was upheld after trial. United States v. Mackey, No. 21-CR-80 (AMD) (SB), 2023 WL 6879613, at *22 (E.D.N.Y. Oct. 17, 2023) (emphasis added). That case does not stand for what the plaintiffs claim. At any rate, it is not clear what false utterances the plaintiffs challenge. They said that the defendants “disseminated wild falsehoods about Who Is Bobby Kennedy” but again, the evidence at this point shows that it was mistakenly and briefly blocked, and is now submitted contested evidence about why the links were blocked. The defendants point to compelling evidence that the video links were incorrectly automatically flagged as a phishing attack, a “not uncommon” response by its automated software to newly created links with high traffic flow. Oppo. 5–6 (citing Mehta Decl. Ex. A ¶ 7). The defendants’ evidence shows that once the defendants were alerted to the problem, through channels set up specifically for that purpose, the links were restored, and the video was made (and is currently still) available on its platform. Mehta Decl. Ex. A. ¶¶ 4–8, Exs. M–Q. Though the plaintiffs say the removal of the video was an effort to coerce them to not urge people to vote for Kennedy, the defendants’ competing evidence shows that it was a technological glitch and that the plaintiffs were aware of this glitch because they reported the problem in the first place. And if the plaintiffs were aware that a tech issue caused the removal of the videos, with that “context” it would probably not be reasonable for them to believe the video links were removed in an effort to coerce or intimidate them. See Wohl III, 661 F. Supp. 3d at 113. Though the evidence may play out differently at a later stage in litigation, at this point I cannot say that this argument or evidence shows serious questions or that the plaintiffs are likely to succeed on their claim. Second, with respect to the warnings received by Reed Kraus and the thirteen-minute suspension of her account, the parties again put forward contested evidence about why the warnings were issued and the account was briefly suspended. The defendants submit evidence from their records showing that Reed Kraus made threats of violence on Meta platforms, entirely unrelated to Kennedy, and that caused her suspension. Mehta Decl. Ex. B ¶¶ 5–10. Their evidence also shows that Reed Kraus’ posts related to Kennedy (and “RFK”) were not affected by any of the defendants’ actions. Id. The plaintiffs tellingly do not contest that Reed Kraus made the threats of violence, instead asserting that the defendants selectively chose to say that “Kennedy” and “RFK” posts were not removed or flagged. See Repl. 10 n.7. But Reed Kraus’ whole theory is that she was targeted for urging people to vote for Kennedy (whose initials, it hardly need be stated, are “RFK”). If her posts relating to Kennedy and RFK were not affected by enforcement, and her suspension arose from her (uncontested) threats of violence—which is what convince a jury that “a reasonable recipient” of Meta’s warnings and suspension would view those communications as any sort of threat to deter her from urging others to vote. See Wohl III, 661 F. Supp. 3d at 113. At this point, with this evidence, the plaintiffs are unlikely to succeed with that argument. Third, the plaintiffs submit evidence that other peoples’ accounts were censored, removed, or threatened with removal when they posted any sort of support for Kennedy and his candidacy. See, e.g., Repl. 1:13–24; [Dkt No. 29-1] Exs. A, B. The defendants fail to respond to these allegations in their opposition, but the reason for this failure seems obvious. Section 11(b) provides a private right of action for Person A where Person B has intimidated, threatened, or coerced Person A “for urging or aiding any person to vote.” 52 U.S.C.A. § 10307(b). It does not on its face, or in any case law I found or the parties cite, provide a private right of action for Person C to sue Person B for intimidating, threatening, or coercing Person A “for urging or aiding any person to vote.” Id. Using that example, the three plaintiffs would be “Person C.” Their evidence very well might suggest that Meta is censoring other users’ pro-Kennedy content.12 But those users are not plaintiffs in this case and are not before me now. Importantly, the plaintiffs had plenty of time and opportunity to add any of those affected users as new plaintiffs in this case, as they added Reed Kraus between filing the initial complaint and filing the AC and current motion. But they did not do so. Nor do they allege or argue that AV24 has some sort of organizational or third-party standing to assert the claims of those affected users.13 And while they seem to say that Kennedy himself is affected because that evidence shows

12 But see [Dkt. No. 40].

13 At least one court reasoned that “the harm protected by Section 11(b) does not readily apply to an organization.” Fair Fight, 2024 WL 24524, at *40. For AV24 to show that it has organizational standing under this statute, it probably must show that the defendants’ actions “have caused it to divert resources” to counteract the defendants’ allegedly unlawful actions. Id. The defendants made several arguments about standing in their papers, which they may reraise in subsequent motions, though for this motion it is enough that Reed Kraus has standing to assert her claims. See Save Bull Trout v. Williams, 51 F.4th 1101, 1106 (9th Cir. 2022) (“[I]n an injunctive case,” the court “need not address standing of each plaintiff if it concludes that one plaintiff has standing.” (citation omitted)); see also LULAC, 2018 WL 3848404, at *2 (discussing the “One Meta users are being coerced or threatened for urging people to vote for him, the effect on the candidate is not what § 11(b) protects. Accordingly, this evidence does not support the plaintiffs’ assertions. The plaintiffs, therefore, fail to counter the compelling evidence and reasons that the defendants identify in explanation for the alleged censorship. More critically, the plaintiffs do not deny the defendants’ portrayal of and reasons for the defendants’ actions. The plaintiffs fail to incorporate those reasons into their assessment of how a “reasonable” recipient of Meta’s communications would interpret the communications in “context.” See Wohl III, 661 F. Supp. 3d at 113. Based on the evidence provided so far, a reasonable recipient of Meta’s communications would be unlikely to view them as even related to voting, let alone as coercing, threatening, or intimidating the recipient with respect to urging others to vote. The plaintiffs fail to draw any parallels with what courts have previously found constituted threats of legal consequences or economic repercussions. The closest allegation is that Reed Kraus lost followers and money when she posted Kennedy-related content.14 Setting aside the above issues with Reed Kraus’ assertions, these alleged impacts share no similarities with those in any other court. They do not have parallel legal consequences (like threats of arrest, deportation, and being sued) or economic consequences (like sending personal information to debt collectors, being evicted, or inducing merchants to refuse to provide necessities). And even if Reed Kraus could show that she lost money or sponsorships after the warnings or suspension, under § 11(b) she would have to show that the losses were due to the warnings or suspension, and not simply due to followers or sponsors opting to leave her page when she posted political content. Cf. Hill v. Williams, No. 16-CV-02627-CMA, 2016 WL 8667798, at *5 (D. Colo. Nov. 4, 2016) (finding plaintiffs had standing to challenge constitutionality of law prohibiting “ballot selfies” in part because some plaintiffs alleged the possibility of social media account termination). Indeed, the plaintiffs do not allege that Meta forced third parties to unfollow Reed Kraus, so her allegation that Meta’s actions somehow led to her loss of 40,000 followers after posting pro-Kennedy content are implausible at best. The far more—perhaps only—reasonable interpretation is that those accounts did not want to see that political content. Accordingly, the plaintiffs fail to show that they are likely to succeed on the merits of their § 11(b) claim because they fail to show that a reasonable recipient of Meta’s communications would view them as coercive, threatening, or intimidating in response to urging others to vote. Because they cannot show this, I need not and do not address the parties’ conflicting arguments about whether the plaintiffs must show the defendants had the specific intent to interfere with their voting rights. 2. Ku Klux Klan Act of 1870, Support or Advocacy Clause, 42 U.S.C. § 1985(3) The plaintiffs assert that the defendants are also violating their rights under the Support or Advocacy Clause of the KKK Act.15 That clause provides in relevant part: [I]f two or more persons conspire to prevent by force, intimidation, or threat, any citizen who is lawfully entitled to vote, from giving his support or advocacy in a legal manner, toward or in favor of the election of any lawfully qualified person as an elector for President . . . ; or to injure any citizen in person or property on account of such support or advocacy . . . whereby another is injured in his person or property, or deprived of having and exercising any right or privilege of a citizen of the United States, the party so injured or deprived may have an action for the recovery of damages occasioned by such injury or deprivation, against any one or more of the conspirators. 42 U.S.C. § 1985(3) (emphasis added). a. Relevant Law The parties discuss the Support or Advocacy Clause without alluding to the complicated and somewhat disparate ways in which cases brought under this clause are treated by courts. Over the last 50 years or so, courts have often conflated the Support or Advocacy Clause with the immediately preceding statutory language of § 1985(3), the Equal Protection Clause.16 These

15 The AC contains a cause of action for violation of 42 U.S.C. § 1986, and the plaintiffs mention it in passing, Mot. 16:4–7, but their motion is not based on that claim.

16 The Equal Protection Clause provides: clauses are distinct, protect different rights, and have separate requirements to establish violations. See, e.g., Kush v. Rutledge, 460 U.S. 719, 720–26 (1983) (articulating some differences between the different clauses). Some of the opinions that confuse or conflate the clauses were cited by the parties here in their papers. To clarify the law regarding the Support or Advocacy Clause, and to properly determine whether the plaintiffs have shown a likelihood of success on or serious questions going to the merits, I start with an overview of the law. Three central questions arise from the opinions and orders addressing the Support or Advocacy Clause: first, whether “racial animus” is required for a claim under this clause, as it is for the preceding Equal Protection Clause; second, whether the clause protects against private action or only state action; and third, whether the clause itself confers a substantive right or if it is instead a remedial clause that provides a remedy for a separate statutory or constitutional violation. First, racial animus is not a required element of a cause of action under the Support or Advocacy Clause. In Griffin v. Breckenridge, 403 U.S. 88, 101–02 (1971), the Supreme Court held that a cause of action under the Equal Protection Clause in § 1985(3) required showing “some racial, or perhaps otherwise class-based, invidiously discriminatory animus behind the conspirators’ action.” The Court based this decision on the statutory language of the Equal Protection Clause that “require[es] intent to deprive of equal protection, or equal privileges and immunities.” Id. at 102. Subsequently, in Kush v. Rutledge, 460 U.S. 719, 720–26 (1983), the Supreme Court explained that the Griffin holding was limited to the Equal Protection Clause. The

or indirectly, any person or class of persons of the equal protection of the laws, or of equal privileges and immunities under the laws; or for the purpose of preventing or hindering the constituted authorities of any State or Territory from giving or securing to all persons within such State or Territory the equal protection of the laws; . . . in any case of conspiracy set forth in this section, if one or more persons engaged therein do, or cause to be done, any act in furtherance of the object of such conspiracy, whereby another is injured in his person or property, or deprived of having and exercising any right or privilege of a citizen of the United States, the party so injured or deprived may have an action for the recovery of damages occasioned by such injury or deprivation, against any one or more of the Court articulated the differences between the various provisions and clauses of § 1985, noting the Griffin opinion relied on the “equal protection” statutory language in the Equal Protection Clause. Id. The Court explicitly held that a showing of racial animus was not required for a claim under § 1985(2) and strongly implied that it would not be required for a claim under the Support or Advocacy Clause or any other clause in § 1985 that did not have the “equal protection” language. See id.; see also United Bhd. of Carpenters & Joiners, Local 610 v. Scott, 463 U.S. 825, 834–39 (1983); Bray v. Alexandria Women’s Health Clinic, 506 U.S. 263, 267–68 (1993) (confirming proof of racial or class-based animus is required for proving “a private conspiracy in violation of the first clause of § 1985(3)” (emphasis added)); Wohl III, 661 F. Supp. 3d at 124 (noting the defendants were “arguing about the wrong clause of the KKK Act in all their submissions” and the Court in Bray “was specifically referring to and analyzing . . . ‘the Equal Protection Clause’”). Because the Support or Advocacy Clause, like § 1985(2), does not have the “equal protection of the law” language, it makes sense that racial animus is not a required element for a cause of action brought under this clause. Cf. Griffin, 403 U.S. at 102; see also LULAC, 2018 WL 3848404, at *5 (reaching same conclusion). Many courts are in accord with this finding. See, e.g., id. at *5–6 (finding the plaintiffs stated a claim under the Support or Advocacy Clause, “which unlike the equal protection part of Section 1985(3) does not require allegations of” racial animus); Wohl III, 661 F. Supp. 3d at 124–25 (“[T]he second clause of 42 U.S.C. Section 1985 (3)—the Support or Advocacy Clause—. . . does not require a showing of racial animus.” (footnote omitted)); Federer v. Gephardt, 363 F.3d 754, 760 (8th Cir. 2004) (“[Plaintiff] is not required to show class-based animus as part of [a] support and advocacy claim.”); Andrews v. D’Souza, 696 F. Supp. 3d 1332, 1345–46 (N.D. Ga. 2023). Second, state action is not a required element for a claim under the Support or Advocacy Clause. In Griffin, 403 U.S. at 96–101, the Supreme Court held that the Equal Protection Clause of § 1985(3) was not limited to state action but also encompassed private conduct. See also Bray, 506 U.S. at 267–68 (providing that “the first clause of § 1985(3)” extends “against private, as well as official” conduct). Subsequently, in United Brotherhood, the Court clarified that the Equal definition a right only against state inference.” 463 U.S. at 833. In that case, the alleged conspiracy was aimed at a violation of the First Amendment, which—as addressed at length above—protects against state abridgment of free speech, not private conduct. It makes sense that, where a plaintiff alleges a conspiracy to deprive her of a right guaranteed by the state, she must allege that the state—the only actor against whom the right applies—was involved in the conspiracy. See also The Support or Advocacy Clause of § 1985(3), 133 Harv. L. Rev. 1382, 1398 (2020) (similar analysis). Recently, the Ninth Circuit cited United Brotherhood for the proposition that “§ 1985(3) requires at least one of the wrongdoers in the alleged conspiracy to be a state actor,” but in that case, as in United Brotherhood, the underlying claim was for violation of the Equal Protection Clause of § 1985(3) based on depriving the plaintiff of their First Amendment rights. Pasadena Republican Club v. W. Just. Ctr., 985 F.3d 1161, 1171 (9th Cir. 2021). In other words, though the Ninth Circuit’s statement appears on its face to apply to all parts of § 1985(3), it is limited to the same boundaries as United Brotherhood; violations of the Equal Protection Clause require at least one conspirator to be a state actor where a First Amendment violation is the underlying right. The Ninth Circuit has not held that the Support or Advocacy Clause of § 1985(3) requires state action, and many district courts have found it does not, allowing cases to proceed against private actors. See, e.g., LULAC, 2018 WL 3848404, at *4–6 (declining to find a state action requirement); Wohl III, 661 F. Supp. 3d at 124–26 (no allegations of state action and no discussion of this being a requirement); Andrews, 696 F. Supp. 3d at 1345–50 (same). I agree with their reasoning for many reasons, including that this clause was part of the KKK Act, which itself was enacted in response at least in part to the threats of the private organization, the KKK, against lawful voters. See, e.g., Support or Advocacy Clause, 133 Harv. L. Rev. at 1389–92 (exploring the history of the act and collecting citations). The court in Cockrum v. Donald J. Trump for President, Inc., 365 F. Supp. 3d 652, 660–65 (E.D. Va. 2019), held otherwise, reasoning that a plaintiff’s claim under the Support or Advocacy Clause failed in the absence of state action. But the court relied entirely on United Brotherhood and Great American Federal Savings & Loan Advocacy Clause, and both of which specifically discussed Griffin, which the Court held in Kush was limited to the Equal Protection Clause. Cockrum, therefore, is unpersuasive. Third, and closely related to the second issue, the Support or Advocacy Clause confers a substantive right. In United Brotherhood, where the Supreme Court addressed only the Equal Protection Clause, it held that “§ 1985(3) . . . ‘provides no substantive rights itself’ to the class conspired against. The rights, privileges, and immunities that § 1985(3) vindicates must be found elsewhere, and here the right claimed to have been infringed has its source in the First Amendment.” United Bhd., 463 U.S. at 833 (quoting Great American, 442 U.S. at 372). The Court did not purport to apply its reasoning to any other clauses in § 1985. Indeed, the opinion consistently reflects and responds to the Court’s prior analysis in Griffin, which explicitly applied only to the Equal Protection Clause. See, e.g., id. at 834 (discussing “the section” of § 1985 addressed in Griffin); Griffin, 403 U.S. at 99, 102 nn.9–10 (repeatedly referring to “the portion of § 1985(3) [now] before us”); id. at 102–03 (discussing the elements of the Equal Protection Clause). Other courts are in accord. See, e.g., LULAC, 2018 WL 3848404, at *5–6 (finding the plaintiffs stated a claim under the Support or Advocacy Clause, “which unlike the equal protection part of Section 1985(3) does not require allegations of . . . violation of a separate substantive right”); Allen, 2021 WL 2223772, at *8 (finding the Support or Advocacy Clause creates a private cause of action (citing 42 U.S.C. § 1985(3)); Wohl I, 498 F. Supp. 3d at 486–87); Ariz. Democratic Party v. Ariz. Republican Party, No. CV-16-03752-PHX-JJT, 2016 WL 8669978, at *5 (D. Ariz. Nov. 4, 2016) (noting the Support or Advocacy Clause provides the injured party “a right of action for recovery of damages”). Again the court in Cockrum, 365 F. Supp. 3d at 661, held otherwise, reasoning that Supreme Court precedent has said that the Support or Advocacy Clause of § 1985(3) “provides no substantive rights itself” and “must be tied to the violation of a substantive constitutional right.” Id. (first citing United Brotherhood, 463 U.S. at 833, 837; and then citing Great American, 442 U.S. at 372). But once again, the two cases upon which the Cockrum court relied concerned the Advocacy Clause. Accordingly, the plaintiffs can establish a violation of their rights under § 1985(3)’s Support or Advocacy Clause without allegations of racial animus, against solely private conduct, and as a standalone violation of their rights. I next address whether they have established a claim. b. The Claim in This Case There are two subsections to the Support or Advocacy Clause. The first looks to whether “two or more persons conspire[d] to prevent by force, intimidation, or threat, any citizen who is lawfully entitled to vote, from giving his support or advocacy in a legal manner, toward or in favor of the election of any lawfully qualified person as an elector for President . . .” 42 U.S.C. § 1985(3). The second looks to whether “two or more persons conspire[d] . . . to injure any citizen in person or property on account of such support or advocacy.” Id. “[T]he elements of a Section 1985(3) claim brought for violations of the Support or Advocacy Clause” under the first subsection are: “(1) a conspiracy; (2) the purpose of which is to force, intimidate, or threaten; (3) an individual legally entitled to vote who is engaging in lawful activity related to voting in federal elections.” Wohl I, 498 F. Supp. 3d at 486–87. The elements of a claim under the second subjection are (1) a conspiracy, and (2) “injuring someone for” advocating for a federal candidate for office. See Andrews, 696 F. Supp. 3d at 1347 (citing 42 U.S.C. § 1985(3)). The plaintiffs fail to show a violation of the first subsection. See Mot. 19:12–20:7; Repl. 8:15–10:12. The definition of “force, intimidation, or threat” is the same in § 11(b) of the VRA as it is in the Support or Advocacy Clause. See Wohl I, 498 F. Supp. 3d at 487; see also Gaetz v. City of Riverside, No. 5:23-CV-01368-HDV-SHKX, 2024 WL 1269311, at *11 (C.D. Cal. Mar. 22, 2024) (“‘Viewed in the light of its origin as a reaction against the “murders, whippings, and beatings committed by rogues in white sheets in the postbellum South,” [Section 1985(3) of] the Ku Klux Klan Act obviously meant to its framers, when it spoke of “force, intimidation, or threat” something much more serious and terrifying’ than tweets and public statements.” (quoting Gill v. Farm Bureau Life Ins. Co. of Mo., 906 F.2d 1265, 1269 (8th Cir. 1990))). As discussed at length Without pointing to that evidence, they do not show likelihood of success on or serious questions going to the merits of the issue. The claim for violation of second subsection also fails, though for different reasons. Under the plain text of the statute, the plaintiffs need not prove that they were forced, intimidated, or threatened by the defendants, but rather that the defendants conspired to injure the plaintiffs or their property, again on account of their advocacy. See 42 U.S.C. § 1985(3); see also Repl. 11:2– 6. The plaintiffs’ motion asserts that they were injured economically when the defendants suspended their accounts because that constituted harm to their property. Mot. 18:22–19:7. But the only plaintiff with a suspended account is Reed Kraus, and the plaintiffs do not contest that her 13-minute suspension was due to her threats of violence rather than her “support or advocacy” for Kennedy. The plaintiffs also do not show economic harm to Reed Kraus or her property based on this suspension. That lack of harm also plagues AV24’s alleged injury—economic and reputational injury from removing links to Who Is Bobby Kennedy—because the record shows that the video was later posted across the defendants’ platforms and separately was seen by millions of people on other platforms. Though it seems possible for the plaintiffs to clarify these injuries to survive a motion to dismiss, the current papers and record have not come close to the required showing of likelihood of success for this motion. The plaintiffs’ final argument under the second subsection of the Support or Advocacy Clause is that Kennedy himself suffered reputational harm and economic injury as a result of the defendants’ censorship, Mot. 20:19–21:3, but this is not quite what the statute protects. To establish a claim, Kennedy must show that the defendants conspired to injure him, and that injury was on account of his advocacy. See Andrews, 696 F. Supp. 3d at 1347 (citing 42 U.S.C. § 1985(3)). The evidence, motions, and AC point only to alleged injury to Kennedy on account of others’ advocacy for him, or injury to others on account of their advocacy. But the statute does not protect a presidential candidate’s right to be supported; it protects the rights of the electorate to support a presidential candidate. Because Kennedy’s asserted injury does not appear to fall within the scope of the statute, the claim is likely to fail on the merits. unlikely to succeed because they do not show they were injured on account of their advocacy. See, e.g., Andrews, 696 F. Supp. 3d at 1347 (noting § 1985(3) prevents conspiracies to prevent someone from “injuring someone for such advocacy”). Once again, the plaintiffs do not contest the evidence that shows Reed Kraus’ suspension was due to threats of violence. The allegations about censoring third parties can only support § 1985(3) claims for those third parties. And though it seems theoretically possible for the plaintiffs to meet this required element of their § 1985(3) claims, at this point they have not shown that they are likely to succeed on—or even that there are serious questions going to the merits of—this question. Because the plaintiffs fail to show a likelihood of success for the above reasons, I need not address their arguments about the intracorporate conspiracy doctrine at this time. 3. The Communications Decency Act, 47 U.S.C. § 230 Finally, the defendants assert that they are immunized against these statutory claims for their editorial decisions to not post the plaintiffs’ content under the Communications Decency Act (“CDA”), 47 U.S.C. § 230. See Oppo. 23:21–25:26. Section 230 of the CDA “protects apps and websites which receive content posted by third-party users (i.e., Facebook, Instagram, . . . etc.) from liability for any of the content posted on their services, even if they take it upon themselves to establish a moderation or filtering system, however imperfect it proves to be.” Est. of Bride by & through Bride v. Yolo Techs., Inc., ---F. 4th---, No. 23-55134, 2024 WL 3894341, at *4 (9th Cir. Aug. 22, 2024). The immunity applies to the decision of “whether to publish . . . third-party content” and whether “to withdraw from publication” such content. Sikhs for Just. “SFJ”, Inc. v. Facebook, Inc., 144 F. Supp. 3d 1088, 1094 (N.D. Cal. 2015), aff’d, 697 F. App’x 526 (9th Cir. 2017) (emphasis added) (quoting Barnes v. Yahoo!, Inc., 570 F.3d 1096, 1102 (9th Cir. 2009), as amended). “This immunity persists unless the service is itself responsible, in whole or in part, for the creation or development of the offending content.” Bride, 2024 WL 3894341, at *4 (quotation marks omitted) (quoting Fair Hous. Council of San Fernando Valley v. Roommates.Com, LLC, 521 F.3d 1157, 1162 (9th Cir. 2008)). asserted by the defendants. Decisions by the defendants to not publish the plaintiffs’ desired content is probably covered by the CDA. See, e.g., Sikhs for Just., 144 F. Supp. 3d at 1094. The plaintiffs seem to assert that the CDA does not apply because they seek to hold the defendants liable for their own speech—their own threats to ban, block, or censor users—not for the speech of others. See Repl. 13:15–14:10. If the plaintiffs could show that the defendants made affirmative “threats” to the plaintiffs, the plaintiffs may be correct that the statements are not covered by the CDA. See Bride, 2024 WL 3894341, at *4. But they have not shown threats, as discussed above. And, this theory of injury seems to contradict other arguments made by the plaintiffs, including that they were injured by the defendants’ refusal to post pro-Kennedy content created by the plaintiffs and third parties. Because this argument is not well fleshed out by the parties, and because it is not dispositive of my findings on this preliminary injunction motion, I decline to address the CDA immunity arguments further in this Order. Should the parties reassert these arguments in future motions, they should clarify the basis for the immunity and the injuries. * * * For those reasons, the plaintiffs fail to establish likelihood of success or serious questions going to the merits of any of their claims.17 The only argument that the plaintiffs make about the remaining Winters factors is that they established all three as a matter of law because they are likely to succeed on the merits of the First Amendment claim. See Mot. 22:24–23:12; Repl. 4:5–12. They make essentially the same argument about their statutory claims, though only in reply. Repl. 4:12–5:9. But they have not shown a likelihood of success or serious questions going to the merits, so their argument fails. I decline to grant the motion on this basis alone. See Leiva-Perez v. Holder, 640 F.3d 962, 966 (9th Cir. 2011) (noting “a certain threshold showing [must be] made on each factor” to be granted a

17 The plaintiffs’ motion to file supplemental evidence in support of their opposition, [Dkt. No. preliminary injunction (citing Cottrell, 632 F.3d at 1131–32)). Additionally, the plaintiffs’ requested injunction seems to be a mandatory injunction, and they do not meet the “high” standard for issuing one. See Doe v. Snyder, 28 F.4th 103, 111 (9th Cir. 2022). “[A] mandatory injunction [is] one that goes beyond simply maintaining the status quo and orders the responsible party to take action pending the determination of the case on its merits.” Id. The plaintiffs request a court “order enjoining Defendants to cease and desist their censorship of speech supporting, advocating for, or urging people to vote for[] Mr. Kennedy.” Repl. 15:7–9. They frame this as a prohibitory injunction to stop censorship, but it seems to be a mandatory injunction to order the defendants to publish the plaintiffs’ posts, videos, speech, and other content about Kennedy. Generally, “mandatory injunctions ‘are not granted unless extreme or very serious damage will result and are not issued in doubtful cases or where the injury complained of is capable of compensation in damages.’” Doe, 28 F.4th at 111 (quoting Marlyn Nutraceuticals, Inc. v. Mucos Pharma GmbH & Co., 571 F.3d 873, 879 (9th Cir. 2009)). The Ninth Circuit has also implied that the “serious questions” sliding scale inquiry for preliminary injunctions does not apply to mandatory injunctions. See id. at 111 n.4. As discussed at length above, the plaintiffs’ case is certainly “doubtful” here. They have not met the likelihood of success or even the serious questions standard. That they request a mandatory injunction to alter the status quo pending litigation of the suit on the merits is all the more reason to deny their request now. I can and will deny it on this basis alone. For the sake of argument and to guide the parties moving forward in this case, below I address why the plaintiffs fail to meet the remaining three Winter factors. A. Irreparable Harm It is not clear from the papers what alleged harm the plaintiffs believe is “irreparable,” in part to their failure to discuss this factor. The Who Is Bobby Kennedy video is available on the defendants’ platforms, and the plaintiffs say that the video has been seen millions of times on other social media platforms. AC ¶ 52. While that does not inherently mean that there is no contrary finding. Reed Kraus’ account is also now active after the 13-minute suspension. To the extent that she asserts permanent damage and irreparable harm going forward based on losing followers after posting pro-Kennedy content, her theory of injury is unreasonable. She says that Meta removed her content and blocked access to her posts; she does not say that Meta removed third party followers from her accounts. As noted, the far more reasonable explanation for losing thousands of followers after posting political content to social media—and only plausible explanation based on the evidence—is that those followers did not want to see that political content. This does not show irreparable harm for the purposes of this motion.18 B. Balance of Equities The balance of equities tips towards the defendants, not the plaintiffs. The defendants’ main argument in their opposition is that the plaintiffs’ requested injunction is barred by the First Amendment. See Oppo. 8:20–10:23. Because violating the defendants’ constitutional rights would be an obvious hardship, I consider this argument under the “balance of equities” analysis of the Winter test.19 Court-ordered injunctions, like all government action, are subject to First Amendment limits. See, e.g., Maldonado v. Morales, 556 F.3d 1037, 1047 (9th Cir. 2009) (citing Madsen v. Women’s Health Center, Inc., 512 U.S. 753, 763 n.2, 764–65 (1994)). A “content-neutral injunction” does not violate the First Amendment so long as the injunction “burden[s] no more speech than necessary to serve a significant government interest.” Id. (quoting Madsen, 512 U.S. at 765). It is hard to see how the plaintiffs’ requested mandatory injunction would be content- 18 Kennedy also ended his presidential campaign and removed his name from the ballots in swing states, after this motion was fully briefed and before the hearing. Elizabeth Findell, Natalie Andrews, Robert F. Kennedy Jr. Drops Out of Presidential Race, Endorses Donald Trump, WALL STREET JOURNAL, Aug. 23, 2024, https://www.wsj.com/politics/elections/robert-f-kennedy-jr- drops-out-of-presidential-race-endorses-trump-f043e9b9, last accessed 8/23/24, 3:15 p.m. Though it does not moot the case, it makes showing irreparable harm harder.

19 In Children’s Defense, 2024 WL 3734422, at *9, the Ninth Circuit briefly discussed Meta’s First Amendment rights in the analysis of the plaintiff’s First Amendment claim, under the second prong of the state action test. Here, the plaintiffs failed to make any argument about the second neutral; they specifically ask me to order the defendants to publish pro-Kennedy campaign content. But the defendants did not brief the issue and so I will apply the less strict standard for content-neutral injunctions to show that even under this lower bar, the plaintiffs fail to meet the standard. The Supreme Court recently held that social media platforms’ decisions and actions to enforce their content moderation policies are protected by the First Amendment. See Moody v. NetChoice, LLC, 144 S. Ct. 2383 (2024). The Court analogized these decisions to editorial and compilation decisions made by newspapers to decide what information to publish and made by parades to decide what floats to allow. See id. at 2399–2403. The Court reasoned that content moderation decisions, including deciding what posts to display and how to order or organize the posts, constituted “expressive activity” akin to the compilation and curation of speech made by these other institutions, and so was protected under the First Amendment—including decisions to exclude speech. Id. at 2401, 2407. Government regulations requiring the publication of other speech in order to “balance” perspectives shared on the platforms, such as the statutes at issue in Moody requiring the social media platforms to allow content that it wanted to exclude, violated the First Amendment rights of the social media companies. See id. at 2407–08.20 Meta’s content moderation decisions to “deprioritize[]” political content are protected by the First Amendment. See id. at 2401. That is true even if it allows supportive speech for all other candidates but Kennedy—“that kind of focused editorial choice packs a peculiarly powerful expressive punch.” Id. at 2402. If I issue an order compelling the defendants here to publish pro- Kennedy content that they have allegedly declined to publish on their platforms, that order will infringe on the defendants’ First Amendment rights. See id.; see also Maldonado, 556 F.3d at 1047. Such an order is constitutionally permissible only if it “burden[s] no more speech than necessary to serve a significant government interest.” Maldonado, 556. F.3d at 1047 (citation 20 The Supreme Court explained that the social media companies brought a facial challenge to the statutes at issue, but the lower courts had analyzed it as an as-applied challenge, so the Supreme Court vacated and remanded the decisions for the lower courts to apply the correct analysis. Moody, 144 S. Ct. at 2397–99, 2409. The Court therefore did not hold these statutes were facially omitted). It is not clear what government interest is served by infringing on Meta’s constitutional rights—the plaintiffs did not brief this, and I can only speculate. A speculative interest, however, is not enough to justify infringement on Meta’s First Amendment rights. And while the plaintiffs argue that threats and intimidation in violation of the VRA § 11(b) are not protected by the First Amendment, as discussed at length above, they fail to establish any of the defendants’ } communication were threats or intimidation. Supra Part 1.B.1. Accordingly, the balance of equities regarding the requested injunction therefore tips sharply in Meta’s favor. C. Public Interest The plaintiffs seem to assert that the public has an interest in hearing from all presidential candidates. But the plaintiffs’ pro-Kennedy content is available on the defendants’ platforms, the content has been widely viewed on other social media platforms, Kennedy is no longer an active candidate, and the public has an interest in ensuring the defendants’ constitutional rights are not violated. It is not clear that this factor favors the plaintiffs. 14 □ □ □ Accordingly, the plaintiffs fail to establish any of the Winter factors favor the entry of a 16 preliminary injunction. CONCLUSION For those reasons, the motion is DENIED. Dated: September 3, 2024 . liam H. Orrick United States District Judge