Pressure on Platforms

The Western District of Louisiana found in July 2023 that officials of the Biden administration likely coerced or significantly encouraged platform moderation (POL-7.11). The Fifth Circuit narrowed the injunction, and the Supreme Court held in 2024 that no plaintiff had standing (POL-7.12). Those findings concern the Biden administration. The record of contact begins earlier.

Special Agent Elvis Chan of the FBI's San Francisco office ran the Bureau's contact with Twitter, Facebook, Google and other companies. He testified on 29 November 2022 that the meetings were quarterly in the 2017 time frame (p. 89). Yoel Roth, then Twitter's head of Site Integrity, stated in a December 2020 declaration that he had held regular meetings since 2018 with ODNI, DHS, the FBI and industry peers (para. 10). The Senate Commerce Committee's majority report dates DHS's Countering Foreign Influence Task Force to March 2018, created for the midterm elections that year (p. 8).

On 4 September 2019 Facebook hosted an election-security meeting with the FBI, DHS and ODNI, and the CISA director attended. Chan testified that the discussion of removing disinformation at that meeting came from the companies (pp. 151 to 154). Scully testified that CISA held regular government-industry meetings with Facebook as the industry lead (p. 21). The 2017 meetings, the 2018 task force and the 2019 meeting fall in the first Trump administration. The Supreme Court's description of meetings before the 2022 midterms falls in the Biden administration.

Chan's answer was: "I would not say it was 100 percent success rate. If I had to characterize it, I would say it was like a 50 percent success rate. But that's just from my recollection." Twitter's brief of 1 June 2023 in Trump v. Twitter states: "The emails were phrased in advisory terms, flagging accounts they believed may violate Twitter's policies—and Twitter employees received them as such, independently reviewing the flagged Tweets." The brief is a litigation position. In footnote 6 Twitter wrote that the bulk of the FBI communications the plaintiffs cited took place before the 2020 elections, while Trump was still in office.

The district court and the Fifth Circuit found that the FBI likely coerced or significantly encouraged the platforms. Those findings were preliminary, and the Supreme Court did not reach the merits (POL-7.11, POL-7.12). The Supreme Court's opinion states that the platforms had "independent incentives to moderate content." The FBI's request and the platform's review are separate acts in each of these documents.

Chan gave his opinion that political pressure from Congress was a contributing factor in the platforms' changes to their terms of service, and he declined to tie the FBI's meetings with congressional staff to account takedowns (p. 132).

CISA, the Cybersecurity and Infrastructure Security Agency, sits within the Department of Homeland Security. Congress passed the CISA Act unanimously, and President Trump signed it on 16 November 2018, according to the Senate Commerce majority report. The agency's election work with platforms is called switchboarding.

Scully testified on 12 January 2023 that "in 2020 we did some switchboard work on behalf of election officials" (p. 16). He defined it as an election official forwarding material the official judged to be disinformation to CISA, which shared it with the appropriate platforms so that they could decide on the content based on their policies (pp. 16 to 17). The House Judiciary majority staff report of 26 June 2023 states: "By 2020, CISA routinely reported social media posts that allegedly spread "disinformation" to social media platforms." The Senate Commerce report gives a Maryland example. A state election board's emailed tweets went through Scully to Twitter, and Twitter replied that the tweets "have been actioned for violations of our policies" (p. 10).

Asked about the period since September 2022, Scully said CISA did not do switchboarding in 2022 (p. 21). The Supreme Court wrote that switchboarding ran until mid-2022 (slip op. p. 5). The House and Senate majority reports say CISA moved election reporting to a nonprofit in 2022. The documents are not reconciled on 2022.

The district court wrote that although CISA's switchboarding started in 2018, no evidence yet produced in the litigation showed that the Trump administration had anything to do with it (ECF p. 118). The Senate Commerce report dates the task force to March 2018 and describes switchboarding in 2020. CISA's position, through Scully, was that switchboarding gave platforms awareness of election officials' reports and that the platforms decided under their own policies (p. 17). The Fifth Circuit held that CISA likely significantly encouraged the platforms, and the Supreme Court did not reach the merits.

On 14 October 2020 Twitter blocked links to the New York Post's reporting on Hunter Biden's laptop, and Facebook demoted the story. The audit of the Twitter Files and Murthy v. Missouri covers the FBI's warnings before that date and the dispute over what the FBI knew about the laptop. Two Twitter documents add who made the decision.

Dorsey wrote that the policy "was informed by conversations with the US government about foreign state misinformation and disinformation and the use of hacked materials or materials of dubious origin being used to manipulate the electorate and influence the outcome of an election." He wrote that the government warnings were repeated before the 2018 and 2020 elections, that Twitter enforced the policy against the Post's links, and that Twitter then narrowed the policy after hearing from journalists.

Roth stated in the declaration that federal law enforcement agencies expected state-actor hack-and-leak operations, likely in October 2020, and that he had learned rumors that such an operation would involve Hunter Biden (para. 11). He stated that Twitter's Trust and Safety leadership determined that the articles violated the Distribution of Hacked Materials Policy and the Private Information Policy and instructed the Site Integrity Team to enforce, which it did at about 10:20 a.m. Pacific on 14 October 2020 (paras. 16 to 17). He stated that neither he nor, to his knowledge, any Twitter employee communicated with Biden for President or the Democratic National Committee about the articles before enforcement (paras. 18 to 19). The declaration is Twitter's statement to a regulator.

Roth told the House Oversight Committee on 8 February 2023 that he had recommended removing amplification and that Twitter instead followed a strict interpretation of the policy. He wrote that "Twitter erred" in acting to avoid repeating the mistakes of 2016.

Meta's chief executive wrote on 26 August 2024 that the FBI warned Meta before the 2020 election about a potential Russian disinformation operation involving the Biden family and Burisma. He wrote that Meta then sent the Post story to fact-checkers and temporarily demoted it. In Twitter's and Meta's own documents, the warning came from the government and the enforcement decision belonged to the company. No document read for this chapter records a government instruction to remove or demote the story.

The letters of Meta and Alphabet reached the House Judiciary Committee, whose Republican majority was investigating the Biden administration. Alphabet's is a statement of facts in 28 numbered paragraphs, submitted in response to subpoenas of 15 February 2023 and 6 March 2025.

Zuckerberg wrote that in 2021 the Biden administration "repeatedly pressured our teams for months to censor certain COVID-19 content, including humor and satire." He wrote: "Ultimately, it was our decision whether or not to take content down, and we own our decisions, including COVID-19-related changes we made to our enforcement in the wake of this pressure." He also wrote that "the government pressure was wrong."

Alphabet's paragraph 8 states: "Senior Biden Administration officials, including White House officials, conducted repeated and sustained outreach to Alphabet and pressed the Company regarding certain user-generated content related to the COVID-19 pandemic that did not violate its policies." The same paragraph states that the company continued to develop and enforce its policies independently.

A White House spokesman responded, as the Agence France-Presse reported on 27 August 2024: "When confronted with a deadly pandemic, this administration encouraged responsible actions to protect public health and safety." The House Judiciary majority staff report of 1 May 2024 reproduces a Meta executive's April 2021 email describing a White House senior adviser who was "outraged" that Meta had not removed a post. The report's statement that the companies changed policies "in ways that were directly responsive to criticism from the Biden Administration" is the Committee's characterization.

Representative Jamie Raskin, the Committee's ranking member, wrote to Alphabet on 30 October 2025 that 20 Alphabet employees interviewed by the Committee "all explained, under penalty of 18 U.S.C. § 1001 for False Statements to Congress, that the Biden Administration never sought to force YouTube to remove content." The transcripts are not public, and the letter is the ranking member's account of them.

The platforms' own reports of what they removed are separate from any request. Meta said on 18 August 2021 that it had removed more than 20 million pieces of COVID-19 and vaccine content and more than 3,000 accounts, Pages and groups. YouTube said on 29 September 2021 that it had removed over 130,000 videos under its COVID-19 vaccine policies. Neither post links a total to a government request. Meta wrote on 7 January 2025 that it had built its content systems "partly in response to societal and political pressure to moderate content," and it ended U.S. third-party fact-checking. Alphabet wrote that YouTube retired its remaining COVID-19 policies by December 2024 and will let creators terminated under them rejoin.

Google, Meta and X each publish counts of government requests.

Google's report says it includes "government requests to review content to determine if it violates our own product community guidelines and content policies." Its U.S. requests fell from 1,819 naming 32,222 items in 2017 to 717 naming 6,991 in 2021. The requester categories were reorganized in 2019, so shares of court orders and agency requests are not comparable across that line. Alphabet's letter of 23 September 2025 describes White House outreach about content that did not violate Google's policies. The report has no row labeled that way, and it does not say whether any White House request is inside the U.S. counts.

No public dataset states how many government reports led to removal. The only figure in the record is the FBI agent's recollection of about half.

State governments act on platform decisions by statute and by flagging. Texas and Florida, both Republican-led, enacted statutes limiting removals. California, Democratic-led, used a flagging office and a reporting statute.

Texas House Bill 20 provides: "A social media platform may not censor a user, a user's expression, or a user's ability to receive the expression of another person based on: (1) the viewpoint of the user or another person" (sec. 143A.002(a)(1)). It covers platforms with more than 50 million active U.S. users a month. Florida's law, approved on 24 May 2021 and effective on 1 July 2021, bars a platform from willfully deplatforming a candidate for office from the date of qualification to the date of the election (s. 106.072(2)). The Supreme Court's opinion quotes Texas's main sponsor describing "West Coast oligarchs" "silenc[ing] conservative viewpoints and ideas," and the Governor's signing remark about silencing conservatives, as the law's stated purpose.

Justice Kagan wrote for the Court: "But a State may not interfere with private actors' speech to advance its own vision of ideological balance." The Court vacated the Eleventh Circuit and Fifth Circuit judgments because neither court did the required facial analysis. On 7 November 2024 the Fifth Circuit remanded the Texas challenge to the district court for fact-intensive questions.

In O'Handley v. Weber the Ninth Circuit recorded that Twitter gave election officials in at least 38 states access to a Partner Support Portal. It quoted the California office's statement that after the 2020 election it flagged for Facebook and Twitter nearly "300 erroneous or misleading social media posts" and that "98 percent of those posts were promptly removed for violating the respective social media compan[ies'] community standards." The court held that even if the office's message was a request to remove the post, compliance "was purely optional," with "no intimation that Twitter would suffer adverse consequences." It held Twitter's moderation was not state action.

California's AB 587 required large platforms to report whether and how they define six categories of content, among them "disinformation or misinformation" and "foreign political interference." X Corp. challenged the content-category reports. On 4 September 2024 the Ninth Circuit held that they "likely fail under strict scrutiny because they are not narrowly tailored." California defended the law as transparency, and the court disagreed on the category reports.

The Digital Services Act is a regulation of the European Parliament and Council of 19 October 2022. Article 9 covers orders from national judicial or administrative authorities against specific illegal content. Article 34(1)(c) lists "any actual or foreseeable negative effects on civic discourse and electoral processes, and public security" among the systemic risks that very large platforms must assess. Article 74(1) caps fines at 6 percent of worldwide annual turnover.

Breton's letter was written "in relation to the planned broadcast on your platform X of a live conversation between a US presidential candidate and yourself." It states: "My services and I will be extremely vigilant to any evidence that points to breaches of the DSA and will not hesitate to make full use of our toolbox, including by adopting interim measures, should it be warranted to protect EU citizens from serious harm." House Judiciary Chairman Jim Jordan replied on 15 August 2024 that the letter was a threat and an intrusion in the American democratic process.

Executive Vice-President Virkkunen answered the Committee on 18 February 2025: "I want to be very clear: the DSA does not regulate speech." The letter states that the Act applies exclusively within the European Union. The Committee's majority staff reports of 25 July 2025 and 3 February 2026 conclude that European regulators use the Act to pressure platforms to change their global moderation rules. Those conclusions are the Committee's. As of 28 May 2026 the Commission's decisions concern transparency, interface design and risk assessment, including its fine of Temu on 28 May 2026 for illegal products.

On 30 August 2024 Justice Alexandre de Moraes of Brazil's Supreme Federal Tribunal ordered X suspended in Brazil until it complied with the Court's orders, paid fines and named a legal representative. The decision directed the agency ANATEL to carry out the suspension and ordered Apple and Google to remove the X app.

Executive Order 13925 directed action about platforms' treatment of users. Executive Order 14149 directs limits on federal officers.

Executive Order 14029, section 1, revokes Executive Order 13925 (Preventing Online Censorship). Section 3(b) of Executive Order 14149 directs the Attorney General to investigate the previous four years and report with recommendations. No report under that directive was found. The consent decree in Missouri v. Biden recites the order in paragraphs 6 to 8.

The decree bars those agencies from threatening Facebook, Instagram, Twitter or X, LinkedIn and YouTube with "some form of punishment" unless they remove or reduce protected speech, and from "unilaterally" directing or vetoing their moderation decisions (paras. 23 to 25). Paragraph 24 states: "This prohibition does not extend to providing Social-Media Companies with information that the companies are free to use as they wish." It permits statements that posts are inaccurate unless coupled with a threat. Paragraph 26 limits the decree to the plaintiffs' own content, and paragraph 17 states that it is not an admission.

The Ninth Circuit held in O'Handley that compliance with the California office's flag "was purely optional" and that there was "no intimation that Twitter would suffer adverse consequences" (p. 16). Twitter's 2023 brief in Trump v. Twitter is a litigation position. It argues that coercion needs the threat of government sanction (pp. 11 to 12) and that the FBI's emails were advisory. The district court and the Fifth Circuit found likely coercion or significant encouragement on the Missouri record (POL-7.11, POL-7.12).