United States v. MacKey

Full title: UNITED STATES OF AMERICA, v. DOUGLASS MACKEY, Defendant.

Court: United States District Court, E.D. New York

Date published: Oct 17, 2023

21-CR-80 (AMD) (SB) (E.D.N.Y. Oct. 17, 2023)

Opinion

21-CR-80 (AMD) (SB)

10-17-2023

UNITED STATES OF AMERICA, v. DOUGLASS MACKEY, Defendant.

ANN M. DONNELLY, United States District Judge:

MEMORANDUM DECISION AND ORDER

ANN M. DONNELLY, United States District Judge:

In seeking relief, the defendant first claims that the government suppressed exculpatory material in violation of Brady v. Maryland, 373 U.S. 83 (1963), when it provided him with FBI interviews of Clinton campaign workers on the second day of testimony. Second, he argues that the government should have disclosed impeachment material about the cooperating witness before Judge Garaufis ruled on the government's application to permit the witness to testify using a pseudonym. Third, he argues that the trial evidence did not establish venue in this district. Finally, he challenges the sufficiency of the evidence in two respects: that the evidence

On March 31, 2023, the defendant was convicted of one count of conspiracy to injure, oppress, threaten or intimidate one or more persons in the free exercise and enjoyment of the right to vote, in violation of Section 241 of Title 18 of the United States Code. (ECF No. 115.) Before the Court are the defendant's motions to set aside the verdict or for a new trial. The Court heard oral argument on September 11, 2023. As explained below, the motion is denied.¹ did not show a “clearly established” violation of Section 241 and that the evidence of conspiracy was legally insufficient. (ECF No. 134.) The government opposes. (ECF No. 140.)

1 This case was originally assigned to Judge Nicholas Garaufis, who ruled on pretrial matters, including the defendant's motion to dismiss the indictment and the government's motion to permit a cooperating witness to testify using a pseudonym. Judge Garaufis contracted COVID-19 after the jury was selected, and the case was reassigned to me on Sunday, March 19, 2023.

BACKGROUND

T. Pretrial Motions

On February 24, 2023, the government sought protective measures to safeguard the identity of its cooperating witness, including allowing him to testify using the name Microchip, which was his online moniker, and to disclose his identity to defense counsel on an “attorney's eyes only” basis. (ECF No. 66 at 1.)² The government also sought to preclude crossexamination about Microchip's cooperation in ongoing, unrelated investigations. (Id.) The defendant opposed, arguing that the jury had to know Microchip's real name to judge “the validity of his testimony.” (ECF No. 73 at 11.) In addition, he maintained that Microchip's cooperation in unrelated matters was relevant to his political or other motivations for testifying, which bore on his credibility. (Id. at 1, 11.)

2 At oral argument on this motion, the government explained that it advised defense counsel of Microchip's identity on February 13, 2021. (Sept. 11, 2023 Oral Argument Transcript (“Sept. 11, 2023 Tr.”) 48:24-49:3.) Defense counsel maintained that he knew Microchip's identity “independent of the government.” (Sept. 11, 2023 Tr. 30:17-18).

Judge Garaufis employed the three-step test described by the New York Court of Appeals in People v. Stanard, 42 N.Y.2d 74, 84 (1977), to determine “whether the government should be ‘permitted to shield a witness's identity, address, or occupation.'” (ECF No. 82 at 6 (quoting United States v. Urena, 8 F.Supp.3d 568, 572 (S.D.N.Y. 2014)).) This test first requires the government to demonstrate why the witness should be permitted to testify anonymously-for example, whether his testimony would subject him to harassment, humiliate him or put him in danger. Urena, 8 F.Supp.3d at 572-73. The burden then shifts to the defendant to demonstrate the materiality of the witness's real name to the issue of guilt. Id. at 573. Relevant factors are: “(1) the extent to which the right to cross-examine is infringed, (2) the relevance of the testimony to the question of guilt or innocence, (3) the nature of the crime charged and the quantum of proof established aside from the testimony of the witness, (4) the nature and significance of the interest or the right asserted by the witness, and (5) the nature of and extent to which the proposed cross-examination would produce evidence favorable to that party and . . . whether such evidence would be merely cumulative.” Id. (quoting Stanard, 42 N.Y.2d at 84). The court then weighs these interests to determine whether the witness's “testimony is sufficiently material to the question of guilt or innocence to overcome the interest to the opposing party.” Id. “In determining materiality, the court is required to keep in mind that the underlying purpose of identity testimony is to establish a background setting in which to test veracity.” Id. (quoting Stanard, 42 N.Y.2d at 84).

Judge Garaufis found that the government made the required showing that disclosure of Microchip's real name could result in online or physical harassment, or danger. (ECF No. 82 at 7-8.) He also concluded that the defense had not met its burden of showing that testimony about Microchip's identity was material to the question of guilt or innocence. (Id. at 8.) Judge Garaufis also noted that allowing Microchip to testify under a pseudonym would “only very slightly” infringe upon the defendant's rights under the Confrontation Clause, because “the relevant exchanges and interactions all took place over the internet, where the CW was known by [the] online moniker[]” under which he would testify at trial, rather than by his legal name or personal background. (Id. at 8.)³

3 Judge Garaufis also granted the government's request to preclude cross-examination about Microchip's ongoing cooperation with law enforcement in other cases, because it fell “outside the bounds of the Government's disclosure requirements” and was not relevant to this case. (ECF No. 82 at 10.) Judge Garaufis gave the defendant leave to renew the argument during trial if he could “articulate a particular basis on which CW's work would be relevant beyond the general proposition that there may be political or other motivations behind the CW's decision to cooperate.” (Id. at 10-11.) The defendant did not renew this argument at trial.

II. The Evidence at Trial

a. The Government's Case

The evidence at trial established the following facts. As of July 2015, the defendant, then 26 years old, began posting on Twitter⁴ under the pseudonym “Ricky Vaughn.”⁵ (Government Exhibit (“GX”) 902.) By September 2016, the defendant, under this pseudonym, had amassed 51,000 followers on Twitter. (GX 200-D at 27.) A research group at the Massachusetts Institute of Technology developed a list, which included politicians and major news outlets, of the 150 “most influential voices” posting about the 2016 presidential election on Twitter. (Trial Transcript (“Tr.”) 258:19-259:3-15, 260:8-10.) The rankings were based on the accounts' influence on Twitter-who was “getting the most retweets”-as well as the extent to which social media commentary poured over into traditional media coverage. (Tr. 259:3-15; GX 1001.) The defendant was ranked 107th. (GX 1001 at 4.)

4 Although Twitter is now known as “X,” the Court uses the platform's name at the time of the conspiracy for ease of reference.

5 The defendant used three Twitter accounts during the conspiracy: @Ricky_Vaughn99, @theRickyVaughn, and @Returnof RV. (GX 902.) These are also known as Twitter “handles.” (Tr. 648:19-22.)

The defendant followed the upcoming presidential election closely and tweeted his observations and commentary about it. On April 10, 2016, the defendant tweeted, “It's a fool's errand to babble on complaining about low-information voters. Either stay out of mass politics or play the game.” (Tr. 401:11-14.) The defendant tweeted many election-related “memes,” which either disparaged Hillary Clinton or supported Donald Trump.

The defendant was also a prominent member of private, invitation-only Twitter direct message groups-the War Room, the Madman Group, and the Micro Chat-of self-described “trolls” whose goal was to develop election memes that would “go viral” and “trend.”⁶ (Tr. 358:18-359:8, 411:17-20, 427:18-23; GX 200-D at 17.) Within these groups, the defendant and other members discussed general “meme-theory”⁷ and strategy to maximize the memes' reach. (See GX 200.) The defendant coordinated retweets and alerted members to hashtags that they should “trend:” for example, taking hashtags associated with pro-Clinton messaging and “hijack[ing]” the hashtags with pro-Trump imagery (GX 200-B at 6; see, e.g., GX 200 at 75 (Ricky Vaughn: “pls help me trend #InTrumpsAmerica”); Tr. 732:21-25), or originating hashtags designed to “cause as much chaos as possible” by creating “controversy . . . for the sole purpose of disparaging Hillary Clinton.” (Tr. 500:3-6, 500:20-24; see, e.g., Tr. 499-500 (discussing the “Podesta emails hashtag”); GX 400 at 28-29 (the War Room admiring the success of the #DraftOurDaughters hashtag).)

6 The defendant was not a member of the Mad Men group or the Micro Chat group during the conspiracy period.

7 The following is an example: “Member: can a meme be anything other than a picture with writing on it?

Ricky Vaughn: yeah definitely[.] meme is any kind of idea that spreads from person to person[.] memes can be represented visually[.]

Member: And with well written tweets? []

Ricky Vaughn: a well written tweet can definitely become a meme[.] An example is Lyin Ted[.] That is now a meme[.] Everyone understands it[.] really good memes go viral[.]

Member: makes sense now[.]

Ricky Vaughn: really really good memes become embedded in our consciousness[.]

Member: Thanks for the input[.]”

(GX 200 at 1-2.)

Microchip,⁸ a member of the conspiracy, testified pursuant to a cooperation agreement.⁹He testified about the terms of the cooperation agreement¹⁰ and about his guilty plea, and the extent to which he stood to benefit by cooperating.¹¹ (See Tr. 559-64.) He described the War Room as a place to share ideas among an exclusive group that included “only people that were considered influential within our circle.” (Tr. 510:14-17.) He explained that his “talent [was] to make things weird and strange so that there is controversy” (Tr. 500:1-3), while the defendant's strengths were amplifying the group's ideas, given his large number of followers, and providing “good ideas for strategies of creating memes, different political messaging” (Tr. 507:9-10). The members of the private Twitter messaging groups also used the websites 4Chan and Reddit,¹²and searched their messaging boards to see “what other content's out there.” (Tr. 498:14-24; GX 410 at 15.) Both Microchip's and the defendant's Twitter accounts had avatars that featured a MAGA hat, which was associated with the Trump campaign. Microchip was not worried that his avatar would undermine the disinformation he tweeted, because those tweets were designed to “spread like wild fire . . . as far as it could go,” eventually reaching the Twitter feeds of users who did not necessarily support Donald Trump. (Tr. 597:18-598:9.)

8 As discussed below, Judge Garaufis permitted the witness to testify using his online moniker.

9 Microchip pled guilty to conspiracy against rights, the same crime for which the defendant was convicted and is awaiting sentencing. (Tr. 480:14-24, 548:7-14.)

10 Microchip testified that he “agreed to possibly be a witness on the case” and to “help [the government] on other cases.” (Tr. 480:17-19.) In exchange for his cooperation, Microchip “get[s] some discounted points . . . on sentencing” and the government “won't charge [him] for other crimes connected to this crime.” (Tr. 481:3-5.) He also hopes that “the [g]overnment will file a letter” on his behalf “with the judge” who will ultimately sentence him. (Tr. at 481:11-13.)

11 The Court granted defense counsel's application to ask Microchip about “his personal interest in remaining anonymous” as long as counsel did not ask questions that would reveal Microchip's identity. (Tr. 473:16-474:22, 476:2-14.) Defense counsel cross-examined Microchip about tweets in which he admitted taking drugs. In one tweet, Microchip said, “3,109 crazy tweets over two weeks. What can I say, I'm insane, [on pills], don't shower, can barely take care of myself, hear voices, talk to the walls, and predict the future.” (Tr. 589:1-6; see Def. Ex. X.) In another, he said, “I'm now 36 hours into my Adderall and ChatGPT marathon” (Tr. 579:24-580:1); another tweet read, “I drink Black Rifle coffee, wear a fishnet trucker hat, have a Jesus tattoo, and inject testosterone” (Tr. at 579:9-13).

12 4chan is an online “messaging board” where people “gather together to talk about different topics.” (Tr. 376:24-25, 497:5-9.) Reddit is also an online messaging board; Microchip testified that it is “kind of like social media, but with “closed-off” groups called “sub-Reddits” that focus on specific topics. (Tr. at 498:2-10.)

Groups like the War Room admired the defendant's influence, which rivaled that of paid media pundits, as well as his success at garnering “impressions.”¹³ (See, e.g., GX 200 at 24, 26, 34.) The defendant considered himself to be a leader of these groups, with particular skill in developing content that could go viral. (GX 200-D at 8 (“I never asked or wanted to be a leader, but so many people are asking it of me, so I feel a responsibility.”); see also Tr. 506:23-507:5, 753:21-754:5 (the defendant acknowledging that his words “carried [weight] with the people in [the Twitter] group”).) By the beginning of 2016, the defendant described his online persona to fellow group members as “powerful. I have something great going on.” (GX 200 at 29.)

13 Twitter defines impressions as “[t]imes a user is served a Tweet in timeline or search results.” X HELP CENTER, Definitions, available at: https://help.twitter.com/en/managing-your-account/using-the-tweet-activity-dashboard.

In the year leading up to the 2016 presidential election, the defendant devoted considerable time and energy to cultivating his social media influence, and described his followers as his “loyal army on Twitter.” (Id. at 32; see also id. at 33 (describing his followers as his “active fans”)). In a May 2016 group message, he reflected on his success: “I feel a giant sense of relief. This sounds dumb but for the past six months I've sacrificed a giant slice of my life to shitpoasting [sic].¹⁴ Looking forward to taking my foot off the gas[.]” (GX 200-D at 13.)

14 The defendant testified that “shitposting” is a Twitter “term of art” that “meant posting a lot of stuff, just kind of [to] distract or get the conversation going, that kind of thing.” (Tr. 666:16-667:2.)

In an October 27, 2016 post, co-conspirator HalleyBorderCol wrote, “[L]et's depress illegal voter turnout [with] a nice hoax.” (GX 400 at 27.) Another group member made a graphic portraying an immigration officer arresting a man who appears to be Latino at a polling place. (Id. at 30.) The same member boasted October 29, 2016 that his meme had “ma[de] the news;” the member attached an article from the Independent online newspaper that characterized the image as “tweeted at Spanish-language outlets likely aimed at intimidating Latino voters.” (Id.) In the group message on the same day, HalleyBorderCol posted images and tweets representing that people could cast their votes by tweeting #Hillary and #PresidentialElection on November 8 (id. at 31-33) or by texting “Hillary” to 59929 (id. at 31). HalleyBorderCol and other group members discussed strategies for maximizing the reach of this misinformation. (See id. at 27, 31-33.)

Microchip saw these images and tweeted on October 30, 2016: “Remember @Hillary Clinton voters, on Nov 8^th, you can vote from home by #Tweeting “#Hillary,” this is only set up for @HillaryClinton voters.” (Id. at 32; see Tr. 509:1-510:24.) Microchip shared the tweet with the War Room. (GX 400 at 32.) Another member of the group responded, “I like that idea- but what if we made it more believable . . . [b]y acting like it's unfair that they can text and vote and we can't[.]” (Id.) Shortly thereafter, someone on Twitter responded, apparently in earnest, to Microchip's “vote from home” tweet: “[W]e should do our own for @realDonaldTrump.” (Id. at 33.) Citing this tweet, Microchip told members of the War Room, “[H]ere's what I worried about, [], people on Trump side thinking this is legit and they stay home, I'm plotting, will have something soon[.]” (Id.) The defendant was a member of the War Room at the time but did not send any messages in this conversation.

On November 1, 2016 at 5:30 p.m., the defendant, still a member of the War Room, published as “Ricky Vaughn” the first of two tweets that announced that people could register their vote by texting on their phones. (Tr. 42:25-43:1, 44:11-19.) The image depicts a Black woman holding a sign that reads “African Americans for Hillary.” (GX 720.)

(Image Omitted)

Transposed on the photograph are two text blocks that read “Avoid the line,” “Vote from home” and “Text ‘Hillary' to 59925. Vote for Hillary and be a part of history.” At the bottom of the image, a “fine print” disclaimer reads, “Must be 18 or older to vote. One vote per person. Must be a legal citizen of the United States. Voting by text not available in Guam, Puerto Rico, Alaska or Hawaii.” In the bottom left corner is an image that resembles the campaign logo-a bold H with a right-pointing arrow. Ricky Vaughn tweeted the image with the hashtags “#ImWithHer” and “#GoHillary.” (Id.)

On November 2, 2016 at 12:30 a.m., the defendant tweeted the second image, depicting a Latina woman sitting in a conference room with a phone and a laptop. (GX 721.)

(Image Omitted)

The image includes text blocks with substantively similar language in Spanish: “Ahorra tiempo. Evita las colas. Vota desde casa o trabajo. Envía un sms escribiendo ‘Hillary' al número 59925 este 8 de Nov.” Another text block reads, “¡Hagamos historias, juntos y juntas! - H.” The image also contains a logo in the bottom left corner and a small-font disclaimer, and Ricky Vaughn's tweet included the same hashtags.¹⁵ (Id.)

15 Robert McNees testified that an apparent typo caused a space between “Go” and “Hillary.” This broke the hashtag so the tweet was only linked to the hashtag “Go.” (Tr. 48:15-23.)

On November 2, 2016, the defendant also retweeted a graphic from War Room member @nia4_trump. (Tr. 505:23-24, 749:16-25; GX 722.)

(Image Omitted)

That graphic has a third image of Hillary Clinton transposed on a blue background and text that read “Save Time Avoid the Line Vote from home,” along with instructions to “Text ‘Hillary' to 59925 and we'll make history together This November 8^th.” The bottom of the graphic contains a small-font disclaimer and H logo. (GX 722.) The accompanying tweet reads, “@TheRickyVaughn thanks for spreading the word! #MAGA #ImWithHer #Vote Hillary from home! Save time & Avoid the line!” (Id.)

Robert McNees, a physics professor and active Twitter user, saw the three tweets, which were similar to graphics he understood to be official Clinton campaign images. (Tr. 53:9-16.) He took screenshots of the images, reported them to Twitter that afternoon, and posted a message to his own social networks encouraging his friends to report them. (Tr. 53:17-54:1.) The same day, Twitter suspended the defendant's account. (Tr. 749:23-25.)¹⁶ On November 4, 2016, the defendant returned to Twitter under a new handle: @Returnof RV. (GX 902.) He tweeted a screenshot of television news coverage of his “voter disinformation attempt”-the tweet containing the image of a Black woman-with the text, “[that feeling when] you haphazardly post a /pol/ meme and it winds up on CNN.” (GX 400 at 39.) The members of the War Room celebrated his return and congratulated him on the reach of the tweet. (Id. at 39-41.)

16 Twitter suspended the defendant's account @RickyVaughn_99 on October 6, 2016. The defendant created a new account, @TheRickyVaughn. That account was suspended on November 2, 2016.

Throughout the next week, members of the Micro Chat group continued to discuss the text-to-vote graphics. (GX 410 at 18-20.) At this point, the defendant was not a member of the Micro Chat group, although there was overlap in membership between the Twitter users in the War Room and Micro Chat. On November 8, 2016, Microchip shared another manufactured image in the Micro Chat-a famous comedian holding a sign saying “Save time. Avoid the line. Vote from home. Tweet ClintonKaine with the hashtag #Presidential Election on November 8th, 2016 between 8 AM and 6 PM to cast your vote.” (Id. at 22.) A member of the chat commented, “I hope some stoners fall for it [] Especially somewhere like Colorado[.]” (Id.)

Jess Morales Rocketto, the digital organizing director of the Clinton campaign, testified that the campaign used text messaging campaigns to “get out the vote” in the days before the election (Tr. 78:1-15) and to answer voters' questions about where, when and how to vote (Tr. 79:17-22.) Rocketto testified that “African American and Latino voters are more likely to use text message than anyone else. Also younger voters love text message, that's the primary way that they communicate.” (Tr. 84:1-6.) People who wanted to “opt-in” to messages from the Clinton campaign could text a short code-47246. (Tr. 82:2-8.)

Rocketto saw the text-to-vote tweets shortly before Election Day. (Tr. 84:13-85:6.) In her view, the graphics were designed to resemble campaign imagery; they incorporated “a really good copy” of the campaign logo, a similar color scheme, fonts and visual style, and disclosures that would be mandated for a paid advertisement. (Tr. 87:14-88:11, 90:6-16, 90:20-23.) She sent the tweets to her bosses, including Theodore Goff, who alerted the campaign's communication team so that they could share the graphics with reporters covering the campaign. (Tr. 91:12-16, 92:23-93:1.)

Lloyd Cotler, the text message campaign manager for the Clinton campaign¹⁷ (Tr. 97:19), was working at the campaign's Brooklyn headquarters when he saw two text-to-vote graphics: “[O]ne was a graphic of an African American woman with a text-your-vote for Hillary call to action on it. And one was in Spanish, with the same kind of language but in Spanish.” (Tr. 99:10-22.)¹⁸ Cotler contacted Mattias Chesley, the Clinton campaign's account manager at Upland Software, the company that provided the text message software platform for the campaign. (Tr. 108:14-20, 109:1-6.) Cotler asked Chesley to identify the operator of the “short code” that the tweets advertised. (Tr. 109:15.)

17 The campaign conducted organizing, fundraising, and voter engagement by text message. (Tr. 97:2225.)

18 On cross-examination, defense counsel asked Cotler about Campaign Employee #1, a campaign worker who monitored social media for the campaign, and his understanding of her job responsibilities. (Tr. 103.)

Chesley alerted Omer Samiri, the president of iVision, which was the company that leased the short code. (Tr. 111:1-3.) Samiri testified that iVision “immediately . . . went on Twitter and started playing some offense to reach out to the Twitter profiles that were sharing this graphic. To basically get them to remove the posts.” (Tr. 118:8-14.) Mr. Samiri knew the ad “wasn't real” (Tr. 119:9), but said that “[t]here was a concern that people would think it was legitimate. I think given technology and innovation and, you know, it won't be outside the norm for something like this, maybe, in the future to exist” (Tr. 120:8-11). In addition, Samiri was worried that iVision would be seen as “somehow involved in this,” and did not want “any punitive action taken on behalf of the carriers to turn off the short code and adversely affect our business.” (Tr. 119:5-6, 119:12-16.) On November 2, 2016, iVision set up an automated response message which appeared when anyone texted the word “Hillary” to the short code; the message informed them that they were not opting into campaign messages, and that they could “text the real Hillary for America for more details.” (Tr. 111:1-22, 119:22-120:2.)

To establish venue, the government called witnesses who testified about the internet infrastructure that connects Twitter users in New York to their online networks. Joel Hendrickson, an engineer for Charter Communications (Tr. 248:16-19), the media company that provides internet services for Spectrum customers (Tr. 248:24-249:1), testified that an internet transmission travels like water, from “a small creek to a river to an ocean” (Tr. 250:22-24). A Spectrum customer connects his device, such as a phone or a laptop, to the Spectrum cable modem in his apartment, either wirelessly or by using an Ethernet cable. (Tr. 250:25-251:5.) The signal from the modem is aggregated with the signals from modems in other apartments, and aggregated again with signals from other apartment buildings, and again, and again, snowballing until it reaches the fiberoptic cables surrounding the island of Manhattan. (Tr. 251:6-14.) In order to leave Manhattan and reach its destination, the signals pass through one of three fiberoptic cables, which run through the Lincoln Tunnel, the Holland Tunnel or the 59th Street Bridge. (Tr. 251:15-20), to Twitter's servers (Tr. 140:2-5). In 2016, Twitter's servers were located in Atlanta and Sacramento. (Tr. 139:25-140:1.) Michael Anderson, an engineer at Twitter, testified that any tweet or private direct message sent by a Twitter user in Manhattan over Spectrum's network would have been routed through both those servers. (Tr. 140:2-141:6.) An FBI agent also provided expert opinion that any signal sent from a Manhattan apartment's home router or a mobile phone data network destined for Atlanta or Sacramento would have to leave the island “through one of these fiber optic lines running under one of the bodies of water surrounding Manhattan.” (Tr. 195:13-15.)

Anderson reviewed the metadata associated with the text-to-vote tweets and determined that the defendant sent the graphics at issue using both his computer and his mobile phone. (Tr. 142:16-143:15.) Marc Bertucci, the defendant's roommate during the conspiracy, testified that their internet provider was Spectrum. (Tr. 147:25.) The venue testimony established that when the defendant tweeted the graphics from his apartment, on devices connected to the internet via Spectrum, those transmissions necessarily traveled through the waters surrounding the island of Manhattan to get to their destination, Twitter's servers in Atlanta or Sacramento.

b. The Defendant's Case

The defendant testified as part of the defense case. In 2014, the defendant, then 24 years old, created a Twitter account. (Tr. 646:6-7.) He estimated that he sent “hundreds of thousands” of tweets over the four years that he was “online as Ricky Vaughn.” (Tr. 646:3-5.) In 2015 and 2016, he used Twitter every day, tweeting or retweeting “[t]ypically hundreds of times per day.” (Tr. 654:23-655:3.) When he began tweeting as Ricky Vaughn, he was employed as an economics analyst, but he was fired in June 2016. (Tr. 652:19-20.) The defendant never revealed his real name to his Twitter followers and had never met any of the members of the War Room or the other private messaging groups in person, including Microchip. (Tr. 646:15-22, 647:4-6.) The defendant's relationships with his followers, including Microchip, was exclusively online; the defendant maintained that the only thing that he and Microchip had in common was their support for Donald Trump. (Tr. 646:23-647:3.) He did, however, address the group members with “terms of endearment,” like “fam” and “team.” (Tr. 687:2-8.) He also referred to them as his friends. (Tr. 739:24-740:5.)

The defendant found the text-to-vote graphics (GX 720, 721) on 4chan (Tr. 647:9-12), and then “copied and pasted” them to Twitter without thought, in “a split second” (Tr. 681:1925.) The next day, @nia4_Trump “mentioned” the defendant in a tweet. (GX 722 (“@TheRickyVaughn thanks for spreading the word!”).) The defendant saw the mention in his notifications and retweeted her message (Tr. 687:12-24),¹⁹ but did not do so to “trick,” “threaten,” “intimidate,” “oppress,” or “injure” voters or their right to vote (Tr. 650:1-12). The defendant estimated that he tweeted or retweeted about 300 times on November 1 and 2, 2016. (Tr. 682:3-7.)

19 Whenever a Twitter user “mentions” another user by writing out his Twitter handle, the mentioned user is automatically notified of the tweet. (Tr. 648:16-649:1.). In this way, Twitter users can direct their public tweets to certain individuals.

The defendant claimed that he did not see the War Room messages in which members workshopped different types of text-to-vote graphics in the days and hours before he tweeted or retweeted the images at ¶ 720, 721 and 722. (Tr. 649:16-25.) Nor did he share the two graphics he claims he found on 4chan (GX 720, 721) with the private Twitter groups, although he did frequently share other memes within those chats or ask its members to photoshop memes for him (Tr. 670:17-671:1, 683:1-5). The defendant belonged to “dozens” of Twitter messaging groups, and could not read every message sent in every group. (Tr. 670:1-13.)

The defendant testified that at “the beginning,” the War Room was a “strategy War Room” and its members aimed to “keep this group open” and “generally free of tweets” so that the members could coordinate with one another. (Tr. 737:4-21; GX 400 at 1.) According to the defendant, however, “it became completely filled with tweets toward the end” and was “basically unreadable;” the defendant claimed that he had notifications for the group “on mute . . . because there were so many messages coming in every day.” (Tr. 737:17-18, 739:7-10.) The defendant did not have to “coordinate” with the members of the War Room because anything “they were talking about” would go “viral” and “trend” and that when his “friends are tweeting about a hashtag and I liked the hashtag, then I will join in.” (Tr. 739:11-740:16.)

The defendant understood “voter turnout” to be “one of the most important things for [political] campaigns” because “it determines whether they win or lose.” (Tr. 672:19-24.) The defendant said that it was not “unusual” for campaigns to “demoralize the opponent's voters.” (Tr. 672:21-673:17.) When he tweeted to his followers that the 2016 election was “on a knife's edge,” he meant that it was “close in the electoral college.” (Tr. 674:20-25.) He told his followers that “the election was close because I believed it, and so they would go vote.” (Tr. 675:1-3.)

The defendant admitted that on November 8, 2016, he retweeted the following tweet from Microchip: “Heard from Hillary they have so many voters out there that if you plan on voting for Hillary just stay home, you're not needed #ElectionDay.” (Tr. 675:12-18, 675:22-24; GX 200 at 123.) The defendant thought the tweet was “funny” and did not “think anyone would take this seriously.” (Tr. 675:20-21.) He only shared the text-to-vote memes to “get publicity,” and was joking when he wrote in the War Room that “Buzzfeed took the bait” by reporting on the memes. (Tr. 678:1-13; GX 410 at 11.) The defendant “thought it was funny that the media thought that this was an attempt to deceive voters, not just a ridiculous post that no one would possibly believe that you could text by vote.” (Tr. 681:1-8.)

The defendant believed that Black voters “were a very important group in the [2016] election,” and that Black voter turnout would be decisive. He tweeted on September 2, 2016 that “the only thing standing in Trump's path is . . . black voters.” (Tr. 702:15-21; GX 200 at 98.) On November 2, 2016, one day after tweeting the vote-to-text meme featuring a Black woman, the defendant tweeted, “Obviously, we can win Pennsylvania. The key is to drive up turnout with non-college whites, and limit black turnout.” (Tr. 702:25-703:4; GX 200 at 105.) The defendant predicted “low black turnout” and tweeted, “Very slight changes in the electorate will lead to a Trump landslide, small increase in white non-college voters, small decrease in blacks.” (Tr. 719:9-13; GX 1005-04). The defendant admitted that he sent multiple tweets, which he characterized as “exaggeration” and “hyperbole,” denigrating Black people. (Tr. 722:13-17; see also Tr. 726:6-15, 727:3-19, 728:9-20.) For example, he described Black people as “gullible,” and wrote, “Black people will believe anything they read, okay Twitter. And we let them vote why?” (Tr. 721:25-722:2, 724:9-10; GX 1005-23.)

He also suggested ways to deter Black people from voting. On September 5, 2016, he tweeted, “Idea: Create #woke #BlackTwitter #nevervote memes, seed them in black social spaces.” (GX 200 at 110.) In another tweet a few minutes later, he offered an example: “A vote for Hillary just means four more years of Hillary Clinton taking black votes for granted. Send her a message, fam. #nevervote.” (GX 200 at 111.)

Nevertheless, the defendant claimed that his selection of a text-to-vote meme depicting a Black woman was “random.” (Tr. 702:4-6.) He agreed that “tricking some number of black people out of voting [would] have served [his] aims overall,” but he “didn't think that meme could actually trick anyone into not voting.” (Tr. 703:7-12.) And, although the defendant believed in 2015 and 2016 that Black people were “stupid” (Tr. 729:9-14), he did not think that “anyone could believe that you could text your vote anonymously. Especially when it's coming from a Twitter account with a MAGA hat ....I don't think it matters whether you're smart, stupid, gullible, or not gullible. I don't think that anyone would fall for that” (Tr. 730:13-20).

At the time of the conspiracy, the defendant also believed that “naturalized citizens should not get the vote.” (Tr. 773:12-17; GX 1005-33; GX 1005-34.) He wrote, “Immigrants, the children of immigrants, et cetera cannot be trusted to vote in the interests of their new country.” (Tr. 774:4-9.) In addition, the defendant believed that women should not be able to vote, a view he expressed with the hashtag “#Repealthe19th.” (Tr. 775:2-16, 776:3-16, 780:20781:4; GX 1005-19 (“Women are children with the right to vote.”); GX 1005-21 (“It's impossible to have a functioning Government when single women and single mothers vote.”).)

The defendant testified that he no longer believed that “black people are stupid,” though he did believe that in 2015 and 2016. (Tr. 729:9-14.) The defendant also testified that he “apologized” “[t]o [his] family” “for the things [he] said as Ricky Vaughn” “[b]ecause it was in bad taste[, i]t was wrong[ and i]t was offensive.” (Tr. 784:14-21.) In April 2018, he began an “in-patient psychotherapy program so [he] could try to turn [his] life around.” (Tr. 687:14688:2.) The defendant explained that he entered the program after “[he] had been doxed in the media” and felt “[he] needed to make a change.” (Tr. 687:16-18.) After two months of the inpatient program, he began receiving out-patient psychotherapy. (Tr. 687:25-688:7.)

III. Midtrial Motions

In his opening statement, defense counsel argued that the text-to-vote “scheme” could not have fooled anyone, and that the timing of the defendant's tweets-a full week before Election Day-refuted the claim that he meant to trick voters. (Tr. 23:3-25:7.) Defense counsel also argued that the government could not prove a conspiracy, because the memes shared in the Twitter messaging groups were not the same memes that the defendant tweeted, and that the defendant shared these memes only “after other memes [] already begun going viral days before.” (Tr. 33:1-7.) Defense counsel maintained that the defendant was merely “shit post[ing]” in order to “[d]istract from the main conversation, get under the skin of the other side, [and] get the other side off message.” (Tr. 33:8-18.)

That same day, eleven witnesses testified as part of the government's case, including Clinton campaign workers Jessica Morales Rocketto and Lloyd Cotler. (See Tr. 216.) As described above, Rocketto was in charge of digital organizing, and she testified about the use of Clinton campaign imagery in the memes, provided context about the campaign's methods of reaching voters by text, and the campaign's reaction to seeing the memes circulate on social media. (See generally Tr. 76-85.) On cross-examination, defense counsel asked Rocketto when Cotler, who reported to her, took action on the memes. (Tr. 95:11-15.) Cotler testified about the campaign's response to text-to-vote disinformation they saw on the internet. (See Tr. 99:10102:3.) Defense counsel asked Cotler when he reached out to the campaign's vendor which administered its own texting short code (47246), and about whether he recalled another employee who monitored social media for the Clinton campaign (“Employee #1”) (Tr. 102:15-103:7); defense counsel also asked Cotler whether Employee #1 specifically monitored 4chan and Reddit²⁰ (Tr. 103:8-18).

20 The Court granted the government's application to seal the names of the non-testifying employees because their privacy interests outweighed the public's interest in disclosure. (Order, Apr. 27, 2023).

The next morning, before testimony began, prosecutors gave defense counsel FBI 302 reports (“302s”) of two interviews with Employee #1. (ECF No. 134 at 18.) The reports reflect Employee #1's concern with the text-to-vote tweets. Employee #1 described the campaign's efforts to get Facebook and Twitter to take the content down. (ECF No. 134-15 at 2.) Employee #1 shared the tweets with the press team, who dealt with legacy media, so that they could alert the general population that it was misinformation. (Id. at 3.) Employee #1 did not recognize the defendant's online moniker or Twitter handle. (Id. at 1.) Employee #1 believed that “[t]here was no one single influencer and no official metric for what [Employee #1] was seeing. These graphics were designed to suppress the vote and began appearing about three months leading up to Election Day.” (Id. at 3-4.) When it came to general misinformation, or “shitposting” on the internet, Employee #1 observed that the digital staffers were overwhelmed by the sheer volume of “shitposting” and misinformation on the internet, and that senior staff did not provide them with enough resources to combat it. (Id. at 3.)

After the election, Employee #1 spoke with lawyers about campaign misinformation in advance of a congressional hearing. (ECF No. 134-14 at 3.) Thereafter, she “experienced harassment from an unknown entity;” she received “strange messages” (id.) and “pictures of [Employee #1] were slid under the door at her residence, along with other ‘creepy stuff.'” (ECF No. 134-15 at 4.) One day, a woman knocked on Employee #1's door and told her that she was being paid to follow Employee #1. (ECF No. 134-14 at 3.) The woman also told Employee #1 that she “would be killed if [she] did not stop reporting on the [social media] misinformation.” (id.; see also ECF No. 134-15 at 4.) The woman visited Employee #1's home “every [] day for five days,” and once yelled at her through her apartment window. (ECF No. 134-14 at 3.)

The defendant moved for a mistrial, claiming that Employee #1's statements to FBI agents were Brady material that should have been produced before trial because they were material to his theories of defense. (Tr. 227:3-10, 227:24-231:23, 234:7-11.) The government explained that it gave the defendant the 302s because defense counsel argued in his opening that his “goal was to rile up the Clinton campaign,” and the campaign's “reactions” to the defendant's tweets “are proof that the campaign was riled up.” (Tr. 629:9-13.) The Court denied the motion without prejudice so that it could review the 302s, the previous day's testimony and the parties' opening statements. (Tr. 234:12-14, 235:2-4.) The government also offered to re-call Cotler and Rocketto for additional cross-examination. (Tr. 236:8-10.)

During the lunch break, at defense counsel's request, the government turned over sixteen more 302s for various campaign employees. (ECF No. 134 at 19; see ECF Nos. 134-3-20.) Three employees discussed the text-to-vote graphics. Employee #2 said that campaign staffers discussed the text-to-vote graphics but “no one knew what to do” next if they identified the “original poster.” (ECF No. 134-4 at 3.) The staffers discussed reporting the tweets to the Department of Justice, but the head of the campaign's communications team told Employee #2 that the campaign should “focus on winning the election,” not the issue of “misinformation around the manner by which voters could cast their vote.” (Id.) Employee #2 believed that the “voter suppression efforts” were “organized” and that those behind the efforts aimed to “disrupt the Clinton Campaign's get out [the] vote efforts.” (Id. at 2-3.) Employee #2 recalled that “there were many memes purporting that people could vote via text message” and thought that these people first coordinated on “private platforms” and then “moved the products to Facebook or Twitter.” (Id. at 3.)

Employee #3, Rocketto's boss, discussed the text-to-vote graphics with her. Although Employee #3 did not recall the specific text-to-vote tweets, he alerted Twitter and Facebook to “advertisements] advising people to vote on the wrong dates.” (ECF No. 134-13 at 1-2.) Twitter removed this content from its platform, but Facebook was resistant. (Id. at 2.) Employee #3 said that “there was a direct concerted effort to coordinate the push of” general misinformation, especially by using the #imwithher hashtag that accompanied the text-to-vote meme. (Id. at 3.)

Employee #4 also monitored social media during the run-up to the 2016 election and recalled seeing text-to-vote memes. (ECF No. 134-20 at 1.) Employee #4 recognized the defendant's online moniker. (Id. at 2.) Employee #4 also said that “the texting tactic was used frequently.” (Id. at 1.) In an effort to “mitigate” the disinformation on Twitter about the time, place and manner of voting, Employee #4 drafted “positive” messaging “re-iterating all the legitimate ways in which someone could vote” (id. at 1-2), while other campaign teams worked directly with Twitter to take down the misleading content (id. at 2). According to Employee #4, junior staff members rang alarm bells about the misinformation, but the senior staff members' response was “lackluster;” Employee #4 observed that things had changed since 2016, and that now “entire teams on campaign staffs” are devoted to dealing with social media disinformation. (ECF No. 134-20.)

Employee #5 worked on the Clinton campaign as a senior social media strategist, focusing mostly on “pushing out the correct voting information” for each state on Facebook. (ECF No. 134-3 at 1, 2.) Employee #5 recalled graphics “mimicking the iconography of the campaign” that contained misinformation like “Did you know you can vote by text” or “vote later than election day.” (Id. at 3.) Employee #5 did not remember the graphics in the defendant's tweets. (Id. at 3-4.) Employee #5 did not recognize the Ricky Vaughn avatar but noted that “there were a lot of trolls, many of whom wore MAGA hats similar to that of [the defendant's] avatar.” (Id. at 4.) Employee #5's team took this very seriously, and it seemed that some “outside sources” mocked them for that.²¹ (Id. at 3.) Employee #5 would have notified Rocketto if there had been a decision about how to react to the disinformation. (Id. at 4.)

21 The report did not include an explanation of what the employee meant by “outside sources.”

The next morning, March 22, 2023, the defendant asked for an adjournment to review the reports and either renew his motion for a mistrial or reopen to the jury and “admit excerpts of the reports into evidence.” (ECF No. 106.) The Court heard testimony for the rest of the day from Microchip and Special Agent Anthony Cunder, the government's summary witness, but granted the defendant's motion for a continuance, adjourning trial for a half-day so that the parties could review the reports and prepare for oral argument on the motion. (Tr. 592:1-8.)

On March 23, 2023, the Court heard argument on the defendant's motion for a mistrial or to reopen to the jury and admit excerpts of the reports. Defense counsel claimed that the 302s for Employees #1 and #4 were Brady material because they showed that the memes were “not a big deal to the Clinton campaign” (Tr. 609:8-12), and that campaign representatives did not report the memes to the Government; according to defense counsel, this information contradicted Rocketto's testimony that the memes were “horrible” and “recognized [] as a problem” (Tr. 610:6-11.) In addition, defense counsel maintained that these 302s, as well as Employee #2's 302, tended to show that the memes were “all over the internet,” including on websites like 4chan. (Tr. 609:12-24.) The defendant claimed that he should have been able to cross-examine Rocketto with this information, because it supported the two theories of defense on which he opened-that the Clinton campaign did not think that the disinformation was a “big deal” supported his argument that the defendant did not intend to trick voters when he tweeted the memes, and that the memes were ubiquitous, making it unlikely that the War Room inspired the defendant to tweet the text-to-vote memes. (Tr. 610:6-612:24.)

The government responded that the campaign's reaction to the defendant's tweets was not probative of the defendant's intent, and that, in any event, the campaign formally responded to the text-to-vote memes in late October.²² (Tr. 624:14-626:4, 628:15-24.) The government also pointed out, in response to defense counsel's arguments about the ubiquity of these memes, that the government had given defense counsel the evidence that text-to-vote memes appeared as early as September 2016, and that it was in fact part of the proof at trial. (Tr. 627:3-20.)

22 The government also explained that it first became aware of the defense claim that his intent was merely to “rile up” the Clinton campaign during defense counsel's opening statement. (Tr. 629:4-13.)

The Court outlined four measures to address counsel's concerns. First, the government could recall Cotler and Rocketto so that defense counsel could cross-examine them about the subjects he raised in his motion for a mistrial. (Tr. 631:2-5.) Second, the government could make the campaign staffers available for the defense to call as part of its direct case. (Tr. 631:67.) Third, the government could call any of the staffers on its case. (Tr. 631:8-9.) Finally, the parties could enter into a stipulation regarding what the staffers would have said. (Tr. 631:1017.) The Court denied the defendant's motion for a mistrial, finding that Clinton campaign staffers' reactions to the vote-to-text memes was neither exculpatory nor material, and rejecting the defendant's argument that the campaign “stand[s] in the shoes of . . . prospective voters” for purposes of Section 241. (Tr. 634:5-23.) The Court asked defense counsel to decide whether he wanted to avail himself of any of the steps the Court suggested. (Tr. 635:2-20.)

After a recess of about 20 minutes (Tr. 635:20-22), defense counsel announced that he did not want to have Cotler or Rocketto recalled for additional cross-examination, nor did he want to call any of the staffers as part of the defense case (Tr. 636:9-17, 637:14-25). Counsel did not ask to interview any of these witnesses to determine whether he wanted to call them as witnesses, nor did he ask that the government call them. The parties eventually agreed to the following stipulation:

In the months prior to the 2016 presidential election, [graphics] containing misinformation concerning how to vote were posted and shared on political messaging boards such as 4chan. At least one member of the Clinton campaign staff observed these [graphics, which] contain[ed] misinformation concerning how to vote on 4chan in the months prior to the 2016 presidential election and brought them to the attention of other staff members after she observed them.

(Tr. 825:6-13; Def. Ex. BB.)²³

23 The defendant wanted the government to stipulate to portions of the 302s without including other aspects of the reports. (ECF No. 134-2.) The government rejected the stipulation as a mischaracterization of the reports.

LEGAL STANDARD

Rule 33 provides that “[u]pon the defendant's motion, the court may vacate any judgment and grant a new trial if the interest of justice so requires.” Fed. R. Crim. P. 33(a). While the court has “broad discretion . . . to set aside a jury verdict and order a new trial to avert a perceived miscarriage of justice,” that discretion should be exercised “sparingly and in the most extraordinary circumstances.” United States v. Ferguson, 246 F.3d 129, 133-34 (2d Cir. 2001) (citation omitted); see also United States v. Gambino, 59 F.3d 353, 364 (2d Cir. 1995) (“Because motions for a new trial are disfavored in this Circuit the standard for granting such a motion is strict.”). “In considering whether to grant a new trial, a district court may itself weigh the evidence and the credibility of witnesses, but in doing so, it must be careful not to usurp the role of the jury.” United States v. Canova, 412 F.3d 331, 348-49 (2d Cir. 2005). The court should grant a Rule 33 motion only if “letting a guilty verdict stand would be a manifest injustice,” because of “a real concern that an innocent person may have been convicted.” Ferguson, 246 F.3d at 134 (citation omitted).

A district court may grant a new trial on the basis of a Brady violation if the late disclosure denied the defendant a fair trial. See, e.g., United States v. Douglas, 525 F.3d 225, 246 (2d Cir. 2008) (“The question [of materiality] is not whether the defendant would more likely than not have received a different verdict with the evidence, but whether in its absence he received a fair trial, understood as a trial resulting in a verdict worthy of confidence.” (quoting Kyles v. Whitley, 514 U.S. 419, 434 (1995))).

Rule 29 provides that “[i]f the jury has returned a guilty verdict, the court may set aside the verdict and enter an acquittal.” Fed. R. Crim. P. 29(c)(1). A court evaluating a Rule 29(c) motion views “the evidence in the light most favorable to the prosecution,” and will uphold the jury's verdict if it determines that “any rational trier of fact could have found the essential elements of the crime beyond a reasonable doubt.” United States v. Facen, 812 F.3d 280, 286 (2d Cir. 2016) (citation omitted).

Viewing the evidence in the light most favorable to the prosecution means “drawing all inferences in the government's favor and deferring to the jury's assessments of the witnesses' credibility.” United States v. Aguiar, 737 F.3d 251, 264 (2d Cir. 2013) (citation omitted). A court “must consider the Government's case in its totality rather than in its parts,” and the sufficiency of the evidence test “may be satisfied by circumstantial evidence alone.” United States v. Wexler, 522 F.3d 194, 207 (2d Cir. 2008) (citations omitted). Thus, a defendant challenging the sufficiency of the evidence “bears a heavy burden.” United States v. Hawkins, 547 F.3d 66, 70 (2d Cir. 2008) (citation omitted).

If the court “enters a judgment of acquittal after a guilty verdict,” it “must also conditionally determine whether any motion for a new trial should be granted if the judgment of acquittal is later vacated or reversed,” and “must specify the reasons for that determination.” Fed. R. Crim. P. 29(d)(1). The standard “that governs whether to grant a new trial under Rule 33” also “applies to determining whether to conditionally grant a new trial under Rule 29(d).” United States v. Full Play Grp., S.A., No. 15-CR-252, 2023 U.S. Dist. LEXIS 155565, at *80 (E.D.N.Y. Sept. 1, 2023) (citing United States v. Finnerty, 474 F.Supp.2d 530, 545 (S.D.N.Y. 2007)).

DISCUSSION

T. Rule 33 Motion for a New Trial

a. Midtrial Disclosures of the 302 Reports

According to the defendant, the government's disclosure of the notes from the interviews of Clinton campaign staffers, on the second day of testimony, violated Brady v. Maryland, 373 U.S. 83 (1963). He claims that the material was exculpatory because it “harmonize[d]” with his defenses that he neither conspired with the members of the War Room nor intended to defraud voters. (ECF No. 134 at 22 (quoting United States v. Mahaffy, 693 F.3d 113, 127 (2d Cir. 2012)); ECF No. 148 at 3.) And although at trial defense counsel said that he had “no issue with these three prosecutors, they've been good to work with” (Tr. 230:3-4), the defendant now accuses the government of intentionally withholding the 302s until the second day of the trial, in a play of “unconstitutional gamesmanship” and “undeniable,” “unfathomable deceit to secure a conviction.” (ECF No. 148 at 10-11, 17-18.) The defendant reiterated this accusation at the hearing on this motion, and claimed that he was not merely “manufacturing indignation.” (Sept. 11, 2023 Oral Argument Transcript (“Sept. 11, 2023 Tr.”) 26:6-7.) He argued that the government's failure to produce the 302s-and the fact that it “withheld” some of Microchip's tweets to Judge Garaufis when applying for Microchip's anonymity-constituted “deliberate” “affirmative misrepresentations or affirmative omissions” that amounted to “prosecutorial fraud.” (Sept. 11, 2023 Tr. 28:21-29:5, 33:12-13.) Under these circumstances, the defendant says, the Court has no other option but to grant his motion. The government responds that the material was not exculpatory, but that in any event, the defendant received the interview reports in time to make use of them at trial.²⁴ (ECF No. 140 at 10, 16-19.)

24 At oral argument on this motion, the government represented that it had “endeavored to turn over things quite early in this case to give [defense counsel] nearly everything we had,” that it had dealt with counsel “in a collegial way,” and that the prosecutors were “shock[ed]” and “dismayed” by the “venom” in defense counsel's accusations. (Sept. 11, 2023 Tr. 63:3-11, 63:19-20.)

As an initial matter, the Court must determine whether the government actually suppressed the 302s for purposes of Brady when it disclosed them on the second day of testimony. If the government did suppress the reports, the question is whether the defendant nevertheless had a reasonable opportunity to use the evidence at trial or to use the evidence to obtain additional evidence for use in the trial. After careful analysis of the evidence at issue and the record as a whole, the Court concludes that the government did not suppress the 302s, that the information contained in them was not exculpatory, and that in any event, the defendant had a reasonable opportunity to use the information during the trial.

The government has a duty to disclose all material evidence favorable to a criminal defendant. United States v. Madori, 419 F.3d 159, 169 (2d Cir. 2005) (citing Brady, 373 U.S. at 87). “Evidence is favorable if it is either exculpatory or impeaching, and it is material if there is a reasonable probability that, had the evidence been disclosed to the defense, the result of the proceeding would have been different.” Mahaffy, 693 F.3d at 127 (citations omitted.) “Where suppressed evidence is inculpatory as well as exculpatory, and ‘its exculpatory character harmonize[s] with the theory of the defense case,' a Brady violation has occurred.” Id. at 13031 (quoting United States v. Triumph Capital Grp., 544 F.3d 149, 164 (2d Cir. 2008)).

“A reasonable probability of a different result is . . . shown when the government's evidentiary suppression undermines confidence in the outcome of the trial.” Douglas, 525 F.3d at 246 (quoting Kyles, 514 U.S. at 434). “Even if evidence is exculpatory and material . . ., the [g]overnment is not required to disclose it ‘if the defendant knows or should have known of the essential facts permitting him to take advantage of any exculpatory evidence.'” United States v. Saipov, 17-CR-722, 2023 U.S. Dist. LEXIS 86030, at *4 (S.D.N.Y. May 16, 2023) (quoting United States v. Frank, 11 F.Supp.2d 322, 327 (S.D.N.Y. 1998)). Rather, the government must “disclose sufficient information to the defendant to ensure that the defendant will not be denied access to exculpatory evidence known only to the Government.” Id.

i. Whether the Evidence is Exculpatory

The defendant does not specify in his written motion which reports he believes contain exculpatory information; he attaches all eighteen to his motion.²⁵ At oral argument, he claimed that seven of the reports were Brady material (Sept. 11, 2023 Tr. 7:17-19), and directed the Court to his proposed stipulation (see supra note 24), which lists the campaign workers whose 302 reports contain purportedly exculpatory material (see ECF No. 134-2).

25 In addition to the reports discussed in the oral motion, the defendant's proposed stipulation included the reports of two campaign workers who documented campaign efforts to research online trends and “shitposting” on Reddit and Twitter. (See ECF No. 134-2.) As discussed below, the defendant does not explain why these campaign efforts were relevant or exculpatory.

The defendant argues that the information in these 302s was exculpatory for two reasons. First, the reports showed generally that the Clinton campaign “recognized that vote-by-text memes were pervasive on multiple places online for months before November 1, 2016” and were “not orchestrated by just one person or group but multiple layers of people,” thereby challenging the government's theory that the defendant conspired with others in private Twitter messaging groups to formulate and workshop these graphics.²⁶ (ECF No. 134 at 24; see also Tr. 622:1922.)

26 The defendant does not identify which reports contained this information; it could be that he is referring to Employee #2, who told agents that he “witnessed countless memes” on 4chan and other websites. (ECF No. 134-4 at 2.)

Second, the reports support the defense that his tweets were mere “shitposting,” intended to “rile up” or distract the Clinton campaign and to make the campaign divert resources that it could have used for other messaging or get-out-the-vote efforts. (ECF No. 134 at 15-17, 21.) The defendant claims that the interview reports, presumably of Employees #1 and #3, reflect that senior campaign officials did not take the misleading tweets seriously enough to take action against them. (Id. at 23-24.) Because the campaign stood in for their voters as “victims” of the conspiracy against rights, the defendant argues, the fact that campaign officials thought the tweets were “not a big deal” supports his defense that he did not intend for the tweets to deceive voters. (Id. at 19; see Tr. 616:25-617:14.)

The government responds that nothing in the 302s is exculpatory; rather, the reports support the government's theory of conspiracy-at least four campaign workers believed that the dissemination of misinformation, including the defendant's tweets, was a coordinated attempt to affect the outcome of the 2016 election, rather than just “shitposting,” and the campaign “did take action to combat the misinformation.” (ECF No. 140 at 10-11, 15-16.) Moreover, the campaign workers' opinions about the nature and significance of the disinformation and shitposting are “wholly irrelevant to the question of the defendant's intent.” (Id. at 13.)

The defendant was charged with violating Section 241 of Title 18 of the United States Code, which is entitled “Conspiracy Against Rights.” The statute provides in relevant part:

If two or more persons conspire to injure, oppress, threaten, or intimidate any person in any State, Territory, Commonwealth, Possession, or District in the free exercise or enjoyment of any right or privilege secured to him by the Constitution or laws of the United States, or because of his having exercised the same ... [t]hey [shall be punished].

The indictment alleged that the objective of the charged conspiracy was to “injure, oppress, threaten or intimidate one or more persons in the free exercise and enjoyment of their right to vote.” (ECF No. 8.) The government had to prove that the defendant knowingly and intentionally joined the conspiracy with the intent to further that objective.

Certainly, the pervasiveness of misleading memes and graphics might support the defense that the defendant “haphazardly” posted content he found in different places on the internet. (See ECF No. 134 at 17, 21, 23-24; ECF No. 134-4 at 2 (Employee #2 stating that he saw disinformation graphics about voting in late September 2016).) But the fact that Clinton campaign staffers recognized that these memes were widespread says nothing about the defendant's specific intent in posting them, or his involvement in orchestrating the spread of the misinformation.

As to the argument that the defendant intended only to rile up the campaign, the reports show, at most, that some senior campaign officials were not as concerned about the defendant's postings and the spread of misinformation as the more junior staffers, who monitored social media and brought the disinformation to their supervisors' attention. This is not exculpatory. Whether the senior members of the Clinton campaign adequately assessed the risk-or had adequate resources to combat the risk-of the defendant's misinformation campaign has nothing to do with whether the defendant knowingly and intentionally conspired to deprive others of the right to vote. The defendant's argument that the campaign and its staffers were “victimized,” as surrogates of their voters, by the conspiracy, is also irrelevant. The defendant's theory is incorrect-it is the voters, not candidates or campaign workers, that the statute was enacted to protect. In any event, the perceptions of the campaign workers-whether they did enough to prevent the defendant from spreading misinformation about how to vote- is not relevant to the defendant's intent.

ii. Whether the Evidence was Material

Evidence is “material” “if there is a reasonable probability that, had the evidence been disclosed to the defense, the result of the proceeding would have been different.” Madori, 419 F.3d at 169 (quoting Pennsylvania v. Ritchie, 480 U.S. 39, 57 (1987)). Materiality is a “resultaffecting test” that requires the prosecutor to predict whether nondisclosure of a certain piece of information would alter the proceeding's outcome, not “an evidentiary test of materiality that can be applied rather easily to any item of evidence ([i.e.,] would this evidence have some tendency to undermine proof of guilt?).” In re United States (Coppa), 267 F.3d 132, 142 (2d Cir. 2001). “[A]s long as a defendant possesses Brady evidence in time for its effective use, the government has not deprived the defendant of due process of law simply because it did not produce the evidence sooner. There is no Brady violation unless there is a reasonable probability that earlier disclosure of the evidence would have produced a different result at trial[.]” Id. at 144 (citations omitted). When “disclosure is first made on the eve of trial, or when trial is under way, the opportunity to use it may be impaired.” Leka v. Portuondo, 257 F.3d 89, 101 (2d Cir. 2001). Accordingly, the Court considers whether the defendant had a “reasonable opportunity to act upon the information efficaciously,” that is, “either to use the evidence in the trial or to use the information to obtain evidence for use in the trial.” United States v. Rodriguez, 496 F.3d 221, 226 (2d Cir. 2007) (citing Leka, 257 F.3d at 100).

As explained above, the Court suggested multiple ways to cure any potential prejudice: 1) recalling the witnesses for cross-examination on the subjects that the defendant claimed were impeachment material; 2) having the government make the staffers available in the event the defendant wished to call them; 3) directing the government to call the staffers; and 4) stipulating with the government about what the witnesses would have said had they testified. The defendant argues that none of these measures would have worked, and that nothing short of a mistrial would suffice.²⁷ However, the defendant rejected three of the suggested measures out-of-hand: recalling Rocketto and Cotler for additional cross-examination, calling the staffers himself, or having the government call them. In fact, defense counsel did not interview the staffers before declining to call them as witnesses.²⁸

27 At oral argument on this motion, the defendant conceded that during trial he asked the court to permit him to re-open and to introduce portions of some 302s; however, he maintained that those steps would have been insufficient. (Sept. 11, 2023 Tr. 6:3-9.)

28 As discussed above, the parties eventually agreed to a stipulation. The defendant faults the government for rejecting his proposal; the government maintained that defense counsel's proposal mischaracterized what the staffers said and omitted significant details. The Court may, in narrow circumstances, force the government to accept a defendant's stipulation to a particular fact “when the point at issue is a defendant's legal status, dependent on some judgment rendered wholly independently of the concrete events of later criminal behavior charged against him,” such as a relevant prior conviction. United States v. Velazquez, 246 F.3d 204, 211 (2d Cir. 2001) (quoting Old Chief v. United States, 519 U.S. 172, 190 (1997)). This case does not fall within those narrow circumstances.

Regardless, the reports that the defendant claims are exculpatory (ECF No. 134-2) would not have lent any meaningful support to the defense. First, staffers' statements about the prevalence throughout the Election of shitposting and voter disinformation (see ECF No. 134 at 21) are immaterial, even if those statements were relevant to the defendant's argument that he was merely participating in that trend to “rile up” the campaign. That the defendant wanted to distract the Clinton campaign does not mean that he could not have also conspired to deprive people of their right to vote; these goals are not mutually exclusive.

Second, statements that senior staffers did not take the posts as seriously as younger, front-line staffers did are also not material to the defense. As discussed above, campaign staffers' perceptions of the posts are irrelevant to the defendant's intent; for the same reason, there is no reasonable probability that, had these statements been disclosed to the defense, the result of the proceeding would have been different. Further, the defendant's claim that the 302s show that the campaign as a whole did not take the memes seriously is not accurate. Employee #3 described the campaign's efforts to work with Twitter and Facebook to remove the text-to-vote graphics from their platform (ECF No. 134-13 at 3). Cotler testified that the campaign decided to share the tweets with media outlets to combat the misinformation. (Tr. 92-93.) Cotler also alerted the company that provided the campaign's text message software platform, which reached out to Omer Samiri at iVision, the company that leased the “short code” used in the defendant's messages. iVision then “started playing some offense” to determine who was behind the false messages, so they could be taken down. (Tr. 118:10-14.) iVision then set up an automated response message to inform people that the defendant's messages were not from the Clinton campaign. (Tr. 119:20-120:2.)

The defendant also argues that if he had the reports before trial, he would have had access to even more “potentially exculpating material,” such as the campaign's daily reports and presentations about 4chan and Twitter, including text-to-vote memes, daily PowerPoint presentations about online trends, and the “records and recollections” of apparently “dozens” of campaign staffers that monitored online misinformation and tracked social media trends. (ECF No. 134 at 24-25.) But the defendant does not explain how any of this information would have led to admissible evidence bearing on the defendant's intent or membership in the conspiracy. Moreover, other information found in the 302s-for example, Employees #1 and #3's opinions that the text-to-vote misinformation was “coordinated”-supported the government's theory of the case. (ECF No. 134-13 at 3; ECF No. 134-15 at 4)

Finally, many of the statements echo information to which the defendant had access well in advance of trial, and therefore do not constitute “evidence known only to the government” that the prosecutors were required to disclose. Saipov, 2023 U.S. Dist. LEXIS 86030, at *10; see also United States v. Payne, 63 F.3d 1200, 1210 (2d Cir. 1995) (“[A] new trial is generally not required when the testimony of the witness is corroborated by other testimony or when the suppressed impeachment evidence merely furnishes an additional basis on which to impeach a witness whose credibility has already been shown to be questionable.” (citations omitted)). For example, before trial, the government produced material that established that text-to-vote graphics began to be disseminated in September 2016; the government turned over an online discussion of text-to-vote disinformation that began on September 26, 2016. (GX 430 at 43.) Similarly, the search warrant affidavit, produced to the defendant in July 2021 and reviewed by the Court, included this information: “In or about and between June 2016 and November 2016, images began appearing on social media websites that purported to be announcements from the presidential election campaign of Hillary Clinton.” (ECF No. 139-1 ¶ 12.) The defendant does not explain why the Rocketto and Cotler interviews, which he concedes were provided to him before trial, were not a sufficient “window” into the “universe” of allegedly exculpatory material from the campaign about the prevalence of “shitposting” and voting misinformation. Indeed, testimony from additional campaign employees would have been largely cumulative of Rocketto's and Cotler's testimony about the campaign's general response to election disinformation, the extent to which the campaign devoted resources to combatting the disinformation, and the specific response to the defendant's postings.

The defendant relies principally on United States v. Mahaffy, 693 F.3d 113 (2d Cir. 2012), in which the defendants “fortuitously found multiple exculpating depositions attended by a member of the prosecution team” only after the defendants were convicted after a retrial. (ECF No. 134 at 21-22.) According to the defendant, this Court is doing what the district court did in Mahaffy: finding “the new information to be much ado about nothing: immaterial, not exculpatory when considered in context, and insufficient to negate evidence of criminal intent.” (Id. at 22 (citing United States v. Mahaffy, et al., No. 05-CR-613, 2010 U.S. Dist. LEXIS 73162 (E.D.N.Y. July 21, 2010), ECF No. 901).) Citing the Second Circuit's observation that a Brady violation occurs when the material's “exculpatory character harmonizes with the theory of the defense case” (id. (quoting Mahaffy, 693 F.3d at 132)), the defendant argues that the 302s in this case are “analogous” (id. at 21), in that they “corroborated and underscored” the defense case that the tweets were mere “shitposting” and that “people were ‘mocked' for taking them seriously” (id. at 23-24.) As discussed above, the defendant contends that the 302s could have led the defense to additional admissible evidence, including campaign materials about online voter misinformation. (Id. at 24-25.)

Mahaffy is an entirely different case, and involved clearly exculpatory material that the government withheld through two trials. 693 F.3d at 133. The Mahaffy defendants were charged with multiple crimes, including securities fraud, securities fraud conspiracy, Travel Act violations, witness tampering, and making false statements. Id. at 121. At the first trial, the jury acquitted the defendants of everything except one count of securities fraud conspiracy, on which the jurors could not reach a verdict, and the trial court declared a mistrial. Id. The defendants were convicted at the second trial. Id. In both trials, the government's theory of securities fraud conspiracy was that the defendants “deprived their employer firms of confidential information.” Id. The “critical issue” was “whether portions of the [transmitted] information actually were confidential.” Id.

The “primary SEC staff attorney investigating the case who conducted almost all of the SEC depositions” was “cross-designated as a Special AUSA” and “sat at counsel table during the trial.” Id. at 122. “Several weeks” before the first trial, this lawyer emailed two members of the prosecution “specifically caution[ing] them that portions of at least one of the deposition transcripts contained possible Brady material[.]” Id. He wrote, “This is the last time I will bring this possible Brady issue up, but look over this excerpt when you get a chance. Tell me whether this requires some disclosure to defense counsel ....” Id. Despite this specific warning from a member of its team, a warning given not for the first time, the government did not disclose the deposition testimony to the defense at either trial. It was not until the SEC administrative proceedings after the second trial that SEC attorneys, who had “close[ly] collaborat[ed]” with the prosecution teams at both trials, disclosed to defense counsel 30 transcripts of investigative depositions, taken before the first trial, that included exculpatory testimony on the “central question” of whether the transmitted information was confidential, as well as testimony that directly “contradicted or undermined the testimony of key government witnesses” on that issue. Id. at 119. There was “no question that the government knew about the transcript that [the SEC attorney] attached, along with the other twenty-nine transcripts, prior to the first trial,” and certainly prior to the second trial. Id. at 122-23.

The Second Circuit concluded that the “withheld SEC testimony strongly suggest[ed]” that the transmitted information was not confidential, as the defense argued, and “called squarely into question the credibility of the government's key cooperating witness.” Id. at 132-33. Therefore, the testimony was “material and favorable to the defense.” Id. at 133. The court also found that there was a “reasonable probability that [the defendants] would not have been convicted had the transcripts been disclosed.” Id. at 134. Accordingly, the court vacated the “component of the defendants' convictions” that relied on the conclusion that the information was confidential. Id.

Unlike Mahaffy, where the government sat on clearly exculpatory material for years and through two trials, the prosecutors in this case turned over the 302s on the second day of the trial. (ECF No. 134 at 21.) The defendant could have called the witnesses himself, recalled the two campaign workers who did testify, or asked the Court to direct the government to call the witnesses. As explained above, the defendant did not do any of those things, and chose not to interview the witnesses.²⁹ Nor, as explained above, do the statements in the 302s bear on a “central question” crucial to the government's theory; the campaign staffers' varying senses of urgency regarding Twitter misinformation has nothing to do with the defendant's intent.

29 The defendant contended at oral argument that it “imposes an unfairness on the defense” to require it to request an adjournment of the trial to take the time required to investigate and interview these witnesses where the defense had already prepared for trial “fastidiously” and given its opening, and where the trial took only “four days,” “the jurors are told it's not going to be more than two weeks,” and the material “should have been turned over before.” (Sept. 11, 2023 Tr. 10:21-12:1.) Defense counsel did not request additional time to interview the witnesses at the trial. As defense counsel conceded, the trial did not take only “four days.” (Sept. 11, 2023 Tr. 20:5-11.) Moreover, the Court never suggested that time would not permit counsel to interview the witnesses.

In short, there is no “reasonable probability” that the outcome of the proceedings would have been different if the government turned over the 302s before trial.

b. The Government's Application to Permit the Cooperating Witness to Testify Anonymously

As explained above, Judge Garaufis concluded that the cooperating witness would face harassment and threats if he were publicly identified, and accordingly granted the government's application to permit the witness to testify as “Microchip.” The defendant argues that Judge Garaufis might have ruled differently had the government disclosed certain information-largely in the form of public tweets-before the application; the defendant cites Microchip's public admissions that he used drugs, owed money to the IRS, feared that public disclosure of his real name would jeopardize his employment, and offered to cooperate with the FBI. (ECF No. 134 at 9.). The defendant concedes that he had all of this information at trial and used it to crossexamine Microchip. (Id. at 9-11.) He claims, however, the government had a Brady obligation to disclose this material, including reports of Microchip's interviews with the FBI, before Judge Garaufis's March 8, 2023 ruling on the government's application, so that the defense could use it to oppose the government's application.³⁰ (Id. at 9-10, 14.) The defendant argues that the government's disclosure of the information on March 10, 2023-two days after Judge Garaufis ruled on the application-was belated. (Id. at 8.) This argument is not persuasive for multiple reasons.³¹

30 As he did in his claims about the interviews of the Clinton staffers, the defendant characterizes the government's disclosures about Microchip as “a cherry-picked sliver of the truth” that render the prosecutors untrustworthy. (ECF No. 148 at 15.) The defendant appears to make this argument to establish a pattern of “deliberate deception” that compels dismissal of the indictment to deter “unconstitutional prosecutorial gamesmanship” (id. at 18), and suggested as much in oral argument (see Sept. 11, 2023 Tr. 39-41). There was, however, no Brady violation.

31 The defendant did not ask Judge Garaufis to reconsider his decision after the government disclosed the impeachment material on Microchip.

First, aside from the fact that the defendant had and used the impeachment material to cross-examine Microchip at the trial, Microchip's tweets cannot support a Rule 33 motion for a new trial based on the government's failure to disclose under Brady because “the evidence . . . was a matter of public record, which [the defendant] himself had the ability to, and ultimately did, locate.” United States v. Moslem, No. 19-CR-457, 2023 U.S. Dist. LEXIS 153310, at *11 (S.D.N.Y. Aug. 30, 2023) (quoting United States v. Teman, 465 F.Supp.3d 277, 343 n.38 (S.D.N.Y. 2020), aff'd, No. 21-1920, 2023 U.S. App. LEXIS 14247 (2d Cir. June 8, 2023)). The government disclosed that Microchip was the confidential informant well in advance of trial on February 13, 2023. As defense counsel stated at oral argument on this motion, he knew Microchip's true identity “independent of the government.” (Sept. 11, 2023 Tr. 30:17-18.) Microchip posted the tweets at issue on his public Twitter account, which the defendant could have found-and did find-at any time after he learned that Microchip was the government's cooperating witness.

The government did not violate its Brady obligation for the same reason. “The [g]overnment does not violate its obligation under those cases where the defendant ‘knew or should have known of the essential facts permitting him to take advantage of any exculpatory evidence.'” Moslem, 2023 U.S. Dist. LEXIS 153310, at *12 (quoting United States v. LeRoy, 687 F.2d 610, 618 (2d Cir. 1982)); see United States v. Zackson, 6 F.3d 911, 918 (2d Cir. 1993) (“[T]he rationale underlying Brady is not to supply a defendant with all the evidence in the Government's possession which might conceivably assist the preparation of his defense, but to assure that the defendant will not be denied access to exculpatory evidence known only to the Government.”); United States v. Constantine, No. 20-4278, 2022 U.S. App. LEXIS 8374, at *2 (2d Cir. Mar. 30, 2022) (summary order) (affirming district court's denial of Rule 33 motion based on alleged Brady violation where defendant had access to facts enabling him to take advantage of exculpatory material); United States v. Whyte, No. 19-CR-64, 2023 U.S. Dist. LEXIS 40465, at *3 (D. Conn. Mar. 10, 2023) (evidence not newly discovered, and no Brady or Giglio violation, where defendant could have found public information regarding witness before trial).

Regardless, Judge Garaufis did not need to evaluate Microchip's general credibility to determine whether he should be permitted to testify using a pseudonym. The “underlying purpose of identity testimony is to establish a background setting in which to test veracity.” Urena, 8 F.Supp.3d at 573 (quoting Stanard, 42 N.Y.2d at 84). To prevail, the defendant would have had to show that Microchip's true name was truly probative with regard to the question of guilt or innocence. As Judge Garaufis found, the defendant did not make that showing, especially since “the relevant exchanges and interactions all took place over the internet,” where Microchip was known by his moniker rather than his “legal name or personal background.” (ECF No. 82 at 8.)

Thus, Microchip's true name did not establish anything about “his credibility or knowledge regarding” the alleged conspirators, especially since the defendant and the alleged coconspirators knew him only by his online moniker. See Urena, 8 F.Supp.3d at 573.³²

32 The defendant argues that it was especially important for the government to turn over impeachment material because Microchip would not have testified if Judge Garaufis denied the government's application. A court considering whether to permit a witness to testify anonymously does not consider whether that witness would testify without a grant of anonymity; rather, it assesses whether restricting cross-examination about the witness's identity denies the defendant his right of cross-examination. See Alford v. United States, 282 U.S. 687, 694 (1931) (“The extent of cross-examination with respect to an appropriate subject of inquiry is within the sound discretion of the trial court.... There is a duty to protect [the witness] from questions which go beyond the bonds of proper cross-examination merely to harass, annoy or humiliate him.”); Smith v. Illinois, 390 U.S. 129, 133-34 (1968) (White, J. concurring) (“I would place in the same category those inquiries which tend to endanger the personal safety of the witness.”) In any event, Microchip testified that his testimony was not contingent on a promise of anonymity. (Tr. 597:14-17.)

II. Motion for Acquittal

a. The Evidence at Trial

The defendant's Rule 29(c) challenge to the sufficiency of the evidence is unavailing.

i. Venue

On May 13, 2022, Judge Garaufis ordered the government to provide a bill of particulars on the issue of venue, and gave the government leave to amend its theory of venue “should that theory change during discovery, motion practice, or while preparing for trial.” (ECF No. 36 at 5 n.1 (citation omitted).) The defendant subsequently moved to dismiss the indictment for improper venue, arguing that the defendant did not commit overt acts in, engage in essential conduct in or have substantial contacts with the district. (ECF No. 54 at 1, 8.)³³ Judge Garaufis determined that the indictment was facially sufficient as to venue on the theory that (1) the defendant's or a co-conspirator's tweeting or retweeting of deceptive images into the Eastern District was an overt act in furtherance of the alleged scheme,³⁴ (2) deceptive images passed through the Eastern District of New York, (3) it was reasonably foreseeable that the defendant's tweets “would reach or pass through a judicial district as large as the Eastern District of New York” on its way to his thousands of followers. (ECF No. 54 at 16-20.) Judge Garaufis rejected the government's effects-based theory that venue would be proper because intended victims of the misinformation conspiracy were “located” in the Eastern District. (Id. at 20-21.)

33 Judge Garaufis's opinion, published on January 23, 2023, is also available at 2023 WL 363595 and 2023 U.S. Dist. LEXIS 11335.

34 Although conspiracy under Section 241 does not require the government to prove an overt act, “venue is proper in any district in which an overt act in furtherance of the conspiracy was committed, even where an overt act is not a required element of the conspiracy offense.” (ECF No. 54 at 13 (quoting Whitfield v. United States, 543 U.S. 209, 218 (2005)).)

In its final instructions, the Court charged that venue in the Eastern District was proper if a preponderance of the evidence established any of the following:

• Either the defendant, or a co-conspirator, or an innocent non-conspirator (caused to act by members of the conspiracy) tweeted an allegedly deceptive image into the Eastern District in furtherance of the alleged scheme, provided that, if tweeted by someone other than the defendant, that act was reasonably foreseeable to the defendant; or

• The allegedly deceptive images sent by the defendant in furtherance of the conspiracy had passed through the Eastern District of New York as they were transmitted to Twitter's servers and beyond, or if deceptive images sent by others in furtherance of the conspiracy had foreseeably passed through the Eastern District of New York as they were transmitted to Twitter's servers and beyond; or

• That the allegedly deceptive images were viewed in the Eastern District and that such viewing (even if innocent) was a foreseeable overt act furthering the ends of the conspiracy.

(ECF No. 114 at 29-30.)

The defendant argues that Judge Garaufis should have granted his pretrial motion to dismiss, and that the Court's instruction that venue can be premised upon the defendant's tweets passing through or being viewed in the Eastern District was erroneous. (ECF No. 134 at 32-33.) The Court declines to disturb Judge Garaufis's holdings concerning the theories by which the government could prove venue at trial on a post-trial motion. That order is the law of the case, and “govern[s] the . . . issue[] in subsequent stages in the same case.” New York v. Adamowicz, 932 F.Supp.2d 340, 345 (E.D.N.Y. 2013) (quoting DiLaura v. Power Auth. of N.Y., 982 F.2d 73, 76 (2d Cir. 1992)). The Court may depart from the law of the case “only when there are ‘cogent' and ‘compelling' reasons for doing so,” which the defendant has not provided. Id. (quoting United States v. Quintieri, 306 F.3d 1217, 1230 (2d Cir. 2002)). Therefore, the Court's instruction to the jury was a correct statement of the law of the case, and was not erroneous.

The defendant also challenges the sufficiency of the evidence of venue, contending that the government did not show that “essential conduct elements” took place in the Eastern District, and thus did not prove venue by a preponderance of the evidence. (ECF No. 134 at 32.) “[W]here the acts constituting the crime and the nature of the crime charged implicate more than one location, venue is properly laid in any of the districts where an essential conduct element of the crime took place.” United States v. Ramirez, 420 F.3d 134, 139 (2d Cir. 2005) (citations omitted); see also 18 U.S.C. § 3237(a) (“Except as otherwise expressly provided . . ., any offense against the United States begun in one district and completed in another, or committed in more than one district, may be . . . prosecuted in any district in which such offense was begun, continued, or completed.”). In this case, the “essential conduct” included the defendant's tweets of text-to-vote graphics to his followers, who he expected would retweet the graphic and “push it [out] further.” (Tr. 753:13-20.)

At trial, the government argued that venue was proper in this District because the deceptive images passed through the internet infrastructure in the Eastern District of New York.³⁵ The defendant lived in Manhattan when he sent the tweets (Tr. 652:24-653:5), and his roommate testified that their internet provider was Spectrum (Tr. 147:25). An engineer from Twitter explained that the defendant sent the tweets using a computer and a phone, and that the tweets were transmitted from his apartment in Manhattan to Twitter's servers in both Sacramento and Atlanta. (Tr. 139-43.) A witness from the FBI and a witness from Spectrum testified that this internet signal would necessarily have passed through fiber optic cables in the waters surrounding the island of Manhattan, on its way to Twitter servers in Sacramento or Atlanta. (Tr. 194:15-25, 250:25-251:20.) Because the Eastern District encompasses the waters surrounding the island of Manhattan, 28 U.S.C. § 112, this testimony established venue by a preponderance of the evidence. See United States v. Brown, 293 Fed.Appx. 826, 829 (2d Cir. 2008) (summary order) (affirming a district court holding that a wire transfer automatically routed through Manhattan was sufficient to find venue in the Southern District); Sept. 3, 2021 Redacted Op. at 46, United States v. Ng Chong Hwa, No. 18-CR-538 (E.D.N.Y. Sept. 10, 2021), ECF No. 84-1 (extending this principle to include use of “Goldman's telecommunications facilities, which transited through the Eastern District of New York”).

35 The government opted to argue venue under this theory, but as discussed above, the Court instructed the jury that the government could prove venue under any of three theories. The defendant proposed that the verdict form include a separate question asking whether the government proved venue (see ECF No. 103); the government objected that it had not “found any examples where the jury has been asked separately on the . . . verdict form[] to find venue” (Tr. 814:8-11; see ECF No. 102). The Court “was not aware of any case law” that required submitting venue to the jury as a separate question (Tr. 822: 8), and accordingly gave the jury a general verdict form. The defendant did not seek a special verdict or interrogatories, pursuant to Rule 49, that would have shown on which theory the jury found venue. Therefore, “the defendant[] waived the ability” to seek, “post-verdict, a new trial based on the failure of [the government] adequately to support one of the three sets of facts on which the general verdict was based.” Morse v. Fusto, 804 F.3d 538, 552 (2d Cir. 2015). In other words, the defense cannot now re-argue the alternative theories of venue that the government opted not to prove after the judge gave all three theories to the jury, and after the defendant did not request a sufficiently detailed special verdict. See id. (finding the defendants “cannot now complain,” after failing to object, “that it is impossible to know whether the verdict would have been different had the inadequately supported [theory] been withdrawn from the jury's consideration”

Judge Garaufis also wrote in his pretrial order that “a reasonable jury could find that logging onto Twitter and viewing the Deceptive Tweets was an overt act and that though the viewers were innocent third parties, their doing so unwittingly furthered the ends of the conspiracy against their right to vote.” (ECF No. 54 at 19 (citing United States v. Royer, 549 F.3d 886, 896 (2d Cir. 2008) (“This includes not just acts by co-conspirators but also acts that the conspirators caused others to take[.]”); United States v. Abdullaev, 761 Fed.Appx. 78, 84 (2d Cir. 2019) (summary order) (“We have repeatedly found venue proper where an out-of-district because allowing them the challenge “would allow them ‘another opportunity to assign as error an allegedly incorrect charge simply because the jury's verdict comports with the trial court's instructions” (quoting Jarvis v. Ford Motor Co., 283 F.3d 33, 56 (2d Cir. 2002))). defendant causes an overt act to be committed by an innocent third party within the district of venue[.]”); United States v. Naranjo, 14 F.3d 145, 147 (2d Cir. 1994) (stating that the defendant “need not have been present in the district, as long as an overt act in furtherance of the conspiracy occurred there”)).) Under this theory, a person who was not a part of or even aware of the conspiracy could still further the conspiracy by retweeting the images; indeed, a key part of the conspirators' strategy was that Twitter users would spread the images “like wild fire” to an even wider population of Twitter users, including those inclined to vote for Hillary Clinton. (See id.; see Tr. 598:1-2.)³⁶

36 In fact, Cotler, one of the Clinton campaign employees, testified that he saw the graphics when they were “going around on social media.” (Tr. 99:10-17.)

Viewing the evidence in the light most favorable to the government, as the Court is required to do, the government established venue in this District by a preponderance of the evidence.

ii. Fair Notice

The defendant argues that acquittal or a new trial is required because “the evidence did not show a violation of Section 241.” (ECF 134 at 29.) In the alternative, he argues that a new trial is warranted “because the jury instructions erroneously adopted the government's overbroad reading of Section 241, defining its scope to forbid anything that may ‘hamper,' ‘frustrate,' ‘slow,' or ‘prevent' (among other things) one's full ability to vote.” (Id. at 31 (quoting ECF No. 114 at 40-41).)

The defendant claims that he did not have fair notice that Section 241 encompassed conduct other than coercion (defined as violence or threats) or ballot-box fraud (defined as destroying votes or stuffing ballot boxes). (Id. at 29.) His conduct, the defendant argues, was “deceit,” which is not encompassed by the word “injure,” and that to read Section 241 to cover conduct “that deceives or discourages the exercise” of the right to vote renders it overbroad under the First Amendment. (Id. at 30.)

As an initial matter, the Court did not instruct the jury that Section 241 covers speech aimed to “deceive” or “discourage” voters. (See ECF No. 114.) In any event, Judge Garaufis held that “the record of historical prosecutions of conspiracies to injure the right to vote, combined with a reasonable reading of the statute itself, constituted ample warning” that the defendant's conduct violated federal law designed to protect voters (ECF No. 54 at 24-30 (collecting election-related § 241 cases), 32-36 (collecting case law defining “injury” within § 241), 36-38 (summarizing DOJ and congressional materials addressing enforcement of § 241)), that the statute's “intent requirement ensures that accidental misinformation will not be criminalized” (id. at 51), and 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” because enforcement will encompass only a “narrow set” of cases (id.). The Court will not disturb that carefully considered ruling on a post-trial motion.

The Court charged the jury that Section 241 “covers conduct intended to obstruct, hinder, prevent, frustrate, make difficult or impossible, or indirectly rather than directly assault free exercise of the right. For example, hinder is defined as to make slow or difficult the progress of, to hamper, to hold back, to prevent, to check.” (ECF No. 114 at 40.)³⁷ The charge was consistent with Judge Garaufis's pretrial decision that Section 241's definition of injury can be caused by acts that “obstruct,” “hinder,” “prevent,” “frustrate,” “make difficult or impossible,” “or indirectly rather than directly assault” the free exercise of a constitutional right. (ECF No. 54 at 32-33) (citing cases defining Section 241 to cover conduct characterized by each of these verbs).)

37 The defendant requested that the Court charge that the statute covers “conduct intended to harm, frighten, prevent, inhibit, or punish the free actions of other persons in the exercise and enjoyment of” the right to vote. (ECF No. 96 at 2.)

iii. Conspiracy

The defendant argues that the government did not prove beyond a reasonable doubt that he participated in a conspiracy. He argues that there was no “meeting of the minds” between the defendant and the other members of the private Twitter messaging groups. (ECF No. 134 at 2728.)

“The government may prove the defendant's knowing participation in a conspiracy through circumstantial evidence,” such as the “defendant's association with conspirators . . . [or] his presence at critical stages of the conspiracy that cannot be explained by happenstance ....” United States v. Anderson, 747 F.3d 51, 60 (2d Cir. 2014) (citations omitted) (collecting cases). “[I]t is enough that the parties have a tacit understanding to carry out the prohibited conduct.” United States v. Nusraty, 867 F.2d 759, 763 (2d Cir. 1989) (quoting United States v. Rubin, 844 F.2d 979, 984 (2d Cir. 1988). “Nevertheless, to be guilty of conspiracy . . . there must ‘be some evidence from which it can reasonably be inferred that the person charged with conspiracy knew of the existence of the scheme alleged in the indictment and knowingly joined and participated in it.'” Id. (citing United States v. Gaviria, 740 F.2d 174, 183 (2d Cir. 1984)).

There was abundant circumstantial evidence from which the jury could infer that the defendant entered into an agreement to deprive others of the right to vote by keeping them from voting in the 2016 presidential election. The government introduced an almost yearlong record of direct messages that established the defendant's membership and affiliation with groups of self-branded “disinformation trolls.” (See, e.g., GX 200-D; GX 400; GX 410.) The members of these groups, including the defendant, coordinated their efforts to spread their political messages and disinformation as widely as possible across Twitter, and ultimately refined their efforts to keep certain groups of voters from voting.

The defendant was not simply present in these groups. To the contrary, he prided himself on his leadership position in the groups, his influence over his many followers-his “loyal army on Twitter”-and on his talent for publicizing memes widely. (GX 200 at 32-33.) The members of the Twitter groups respected him and saw him as a leader. (Tr. 428:11-17, 506:25-507:1.) The defendant sometimes participated in the groups' strategy discussions; even when he did not join in the conversation, he regularly tweeted out the political messages or disinformation that the groups decided to push. (See, e.g., Tr. 842-50.)

As the presidential election neared, the evidence showed that these groups increasingly focused on depressing voter turnout by spreading disinformation about the time, place and manner of the election. In the days leading up to the defendant's posts, the groups strategized about how to design their disinformation-in the form of false text-to-vote ads-so that Hillary Clinton voters would believe they were real. They also considered when the graphics should be distributed-either immediately before the election or somewhat earlier, so that early voting might be affected-and whether the graphics were effective. (See, e.g., Tr. 376:3-14, 459:11-14, 511:10-25, 512:15-513:4.)

On October 27, 2016, HalleyBorderCol suggested in the War Room that they “depress illegal voter turnout [with] a nice hoax.” (GX 400 at 27.) Another group member boasted that his graphic portraying an immigration officer arresting a Latino man at a polling place “ma[de] the news” and attached an article saying that the image was “likely aimed at intimidating Latino voters.” (Id. at 30.) In the War Room on the same day, HalleyBorderCol posted vote-by-tweet and vote-by-text images and discussed strategies for maximizing the reach of this misinformation. (Id. at 31.) On October 28 and 29, 2016, Jakekass and 1080p discussed “work[ing] on a new fake ad campaign.” (Tr. 460:22-461:17.)

Two days before the defendant made his posts, Microchip shared this tweet with the War Room: “Remember @Hillary Clinton voters, on Nov 8^th, you can vote from home by #Tweeting “#Hillary,” this is only set up for @HillaryClinton voters.” (Tr. 510; GX 400 at 32.) Another member of the group wanted to “make it more believable” and to “give the impression that only Hillary Clinton voters can do this,” not Trump voters. (Tr. 511:15-23.) Then, on November 1 and 2, 2016, the defendant tweeted the two false text-to-vote graphics and retweeted a third, all of which pushed the same message as the tweets by HalleyBorderCol and Microchip, and were similarly focused on Hillary Clinton voters. (Tr. 386:2-11; GX 720; GX 721; GX 722.) The defendant was suspended from Twitter for tweeting the false ads but returned shortly thereafter; the group members celebrated his return and congratulated him on the tweets. (Tr. 749:24.)

Moreover, the defendant's efforts to affect the election-by discussing, creating and spreading deceptive information about how to vote-also supported the conclusion that his tweets were not mere “coincidence” unrelated to the planning in the War Room in the days and weeks before the election. Compare Nusraty, 867 F.2d at 762. The defendant repeatedly tweeted about how close the election would be (Tr. 703:20-704:1) and was focused on “turn[ing] out” groups of voters that supported Trump and limiting turnout of those who did not, like Black voters (Tr. 702:25-703:4). He also posted other tweets leading up to the election, beyond the false text-to-vote ads, that were aimed at encouraging Black people not to vote. (Tr. 704:2-8.) He discussed that strategy with members of the Twitter messaging groups. (Tr. 703:4-6.) The defendant's messages with his co-conspirators and his tweets throughout the campaign illustrate that he shared the same objective as the members of the conspiracy-preventing Hillary Clinton supporters from voting-and he tweeted the fraudulent text-to-vote graphics as part of that effort.

The Court has “examined the entire case, take[n] into account all facts and circumstances, and [made] an objective evaluation.” Aguiar, 737 F.3d at 264 (quoting Ferguson, 246 F.3d at 134). “[V]iewing the evidence in the light most favorable to the prosecution,” “deferring to the jury's assessment of the witnesses' credibility,” and considering the evidence “in its totality, not in isolation,” the Court finds that the defendant has not sustained his “heavy burden” to show that he is entitled to relief under Rule 29. Id. A rational jury could conclude from all the evidence that the defendant, together with others, conspired “to injure, oppress, threaten, or intimidate” one or more people “in the free exercise or enjoyment of any right or privilege secured to him by the Constitution or laws of the United States:” specifically, the right to vote. 18 U.S.C. § 241. For these reasons, the defendant's motion is denied.

CONCLUSION

For these reasons, the defendant's motion for a new trial or acquittal is denied.

SO ORDERED.