The following is adapted from a talk delivered at Hillsdale College on February 7, 2023.

Elon Musk’s takeover of Twitter last October and the subsequent reporting on the Twitter Files by journalists Matt Taibbi, Bari Weiss, and a handful of others beginning in early December is one of the most important news stories of our time. The Twitter Files story encompasses, and to a large extent connects, every major political scandal of the Trump-Biden era. Put simply, the Twitter Files reveal an unholy alliance between Big Tech and the deep state designed to throttle free speech and maintain an official narrative through censorship and propaganda. This should not just disturb us, it should also prod us to action in defense of the First Amendment, free and fair elections, and indeed our country.

After Musk completed his acquisition of Twitter, he fired a slew of useless or insubordinate employees, instituted new content moderation policies, and tried to reform a woke corporate culture that bordered (and still borders) on parody. In the process, Musk coordinated with Taibbi and Weiss on the publication of a series of stories based on internal Twitter documents related to an array of major political events going back years: the Hunter Biden laptop scandal, Twitter’s secret policy of shadow banning, President Trump’s suspension from Twitter after the January 6 U.S. Capitol riot, the co-opting of Twitter by the FBI to suppress “election disinformation” ahead of the 2020 election, Twitter’s involvement in a Pentagon overseas psy-op campaign, its silencing of dissent from the official Covid narrative, its complicity in the Russiagate hoax, and its gradual capitulation to the direct involvement of the U.S. intelligence community—with the FBI as a go-between—in content moderation.

As Taibbi has written, the Twitter Files “show the FBI acting as doorman to a vast program of social media surveillance and censorship, encompassing agencies across the federal government—from the State Department to the Pentagon to the CIA.”

The Twitter Files contain multitudes, but for the sake of brevity let us consider just three installments and their related implications: the suppression of the Hunter Biden laptop story, the suspension of Trump, and the deputization of Twitter by the FBI. Together, these stories reveal not just a social media company willing to do the bidding of an out-of-control federal bureaucracy, but a federal bureaucracy openly hostile to the First Amendment.

Hunter Biden’s Laptop

On October 14, 2020, the New York Post published its first major exposé based on the contents of Hunter Biden’s laptop, which had been dropped off at a Delaware computer repair shop in April 2019 and never picked up. It was the first of several stories detailing Biden family corruption and revealing the close involvement of Joe Biden in his son’s foreign business ventures in the years during and after Biden’s vice presidency. Hunter, although doing no real work, was making tens of millions of dollars from foreign companies in places like Ukraine and China. The Post’s bombshell reporting shined a bright light on what was happening.

According to the emails on the laptop, Hunter introduced then-Vice President Biden to a top executive at Burisma, a Ukrainian energy company that was paying Hunter (who had no credentials or experience in the energy business) up to $50,000 a month to sit on its board. Soon after this meeting, Vice President Biden pressured the Ukrainian government to fire a prosecutor investigating the company. In an earlier email, a top Burisma executive asked Hunter for “advice on how you could use your influence” to benefit the company. The Post’s ensuing stories revealed more of the same: a shocking level of corruption and influence-peddling by Hunter Biden, whose emails suggest his father was closely connected to his overseas business ventures. Indeed, those ventures appear to consist entirely of Hunter providing access to Joe Biden.

Twitter did everything in its power to suppress the Biden story. It removed links to the Post’s reporting, appended warnings that they might be “unsafe,” and prevented users from sharing them via direct message—a restriction previously reserved for child pornography and other extreme cases. In an extraordinary step, Twitter also locked the Post’s account and the accounts of anyone who shared links to its reporting, including White House Press Secretary Kayleigh McEnany. These actions were justified under the pretext that the stories violated Twitter’s hacked-materials policy, even though there was no evidence, then or now, that anything on the laptop was hacked.

Twitter executives at the highest levels were directly involved in these decisions. Former head of Legal, Policy, and Trust Vijaya Gadde, the company’s chief censor, played a key role, as did former head of Trust and Safety Yoel Roth. Oddly, all this seems to have been done without the knowledge of Twitter’s then-CEO Jack Dorsey. And it was done despite internal pushback from other departments.

“I’m struggling to understand the policy basis for marking this as unsafe,” wrote a Twitter communications executive in an email to Gadde and Roth. “Can we truthfully claim that this is part of the policy?” asked former VP of Global Communications Brandon Borman. His question was answered by Deputy General Counsel Jim Baker—a former top lawyer for the FBI and the most powerful member of a growing cadre of former FBI employees working at Twitter—who said that “caution is warranted” and that some facts “indicate the materials may have been hacked.”

But there were no such facts, as Baker and other top Twitter executives knew at the time. The laptop was exactly what the Post said it was, and every fact the Post reported was accurate. Other major media outlets like The New York Times and The Washington Post would begrudgingly admit as much 18 months later, after Joe Biden was ensconced in the White House.

If there were no hacked materials in the Post’s reporting, why did Twitter immediately react as if there were? Because long before the Post published its first laptop story, there had been an organized effort by the intelligence community to discredit leaked information about Hunter Biden. The laptop, after all, had been in federal custody since the previous December, when the FBI seized it from the computer repair shop. So the FBI knew very well that it contained evidence of straightforward criminal activity (such as illicit drug use) as well as of corruption and influence-peddling.

The evening before the Post ran its first story on the laptop, FBI Special Agent Elvis Chan sent ten documents to Roth at Twitter through a special one-way communications channel the FBI had established with the company. For months, the FBI and other federal intelligence agencies had been priming Roth to dismiss news reports about Hunter Biden ahead of the 2020 election as “hack-and-leak” operations by state actors. They had done the same thing with Facebook, whose CEO Mark Zuckerberg admitted as much to Joe Rogan in an August 2022 podcast. As Michael Shellenberger reported in the seventh installment of the Twitter Files, the FBI repeatedly asked Roth and others at Twitter about foreign influence operations on the platform and were repeatedly told there were none of any significance. The FBI also routinely pressured Twitter to hand over data outside the normal search warrant process, which Twitter at first resisted.

In July 2020, Chan arranged for Twitter executives to get top secret security clearances so the FBI could share intelligence about possible threats to the upcoming presidential election. The next month, Chan sent Roth information about a Russian hacking group called APT28. Roth later said that when the Post’s story about Hunter Biden’s laptop broke, “It set off every single one of my finely tuned APT28 hack-and-leak campaign alarm bells.” Even though there was never any evidence that anything on the laptop was hacked, Roth reacted to it just as the FBI had conditioned him to do, using the company’s hacked-materials policy to suppress the story as soon as it appeared, just as the agency suggested it would, less than a month before the election.

Suspending the President

The erosion of Twitter’s content moderation standards would continue after the Hunter Biden laptop scandal, reaching its apogee on January 8, 2021, two days after the Capitol riot. That is when Twitter made the extraordinary decision to suspend President Trump, even though he had not violated any Twitter policies. As the Twitter Files show, the suspension came amid ongoing interactions with federal agencies—interactions that were increasing in frequency in the months leading up to the 2020 election, during which Roth was meeting weekly with the FBI, the Department of Homeland Security, and the Office of the Director of National Intelligence. As the election neared, Twitter’s unevenly applied, rules-based content moderation policies would steadily deteriorate.

Content moderation on Twitter had always been an unstable mix of automatic enforcement of rules and subjective interventions by top executives, most of whom used Twitter’s censorship tools to diminish the reach of Trump and others on the right through shadow banning and other means. But that was changing. As Taibbi wrote in the third installment of the Twitter Files: “As the election approached, senior executives—perhaps under pressure from federal agencies, with whom they met more as time progressed—increasingly struggled with rules, and began to speak of ‘vios’ [violations] as pretexts to do what they’d likely have done anyway.”

After January 6, Twitter jettisoned even the appearance of a rules-based moderation policy, suspending Trump for a pair of tweets that top executives falsely claimed were violations of Twitter’s terms of service. The first, sent early in the morning on January 8, stated: “The 75,000,000 great American Patriots who voted for me, AMERICA FIRST, and MAKE AMERICA GREAT AGAIN, will have a GIANT VOICE long into the future. They will not be disrespected or treated unfairly in any way, shape or form!!!” The second, sent about an hour later, simply stated that Trump would not be attending Joe Biden’s inauguration on January 20.

That same day, key Twitter staffers correctly determined that Trump’s tweets did not constitute incitement of violence or violate any other Twitter policies. But pressure kept building from people like Gadde, who wanted to know whether the tweets amounted to “coded incitement to further violence.” Some suggested that Trump’s first tweet might have violated the company’s policy on the glorification of violence. Internal discussions then took an even more bizarre turn. Members of Twitter’s “scaled enforcement team” reportedly viewed Trump “as the leader of a terrorist group responsible for violence/deaths comparable to Christchurch shooter or Hitler and on that basis and on the totality of his Tweets, he should be de-platformed.”

Later on the afternoon of January 8, Twitter announced Trump’s permanent suspension “due to the risk of further incitement of violence”—a nonsense phrase that corresponded to no written Twitter policy. The suspension of a sitting head of state was unprecedented. Twitter had never taken such a step, even with heads of state in Nigeria and Ethiopia who actually had incited violence. Internal deliberations unveiled by the Twitter Files show that Trump’s suspension was partly justified based on the “overall context and narrative” of Trump’s words and actions—as one executive put it—“over the course of the election and frankly last 4+ years.”

That is, it was not anything Trump said or did; it was that Twitter’s censors wanted to blame the President for everything that happened on January 6 and remove him from the platform. To do that, they were willing to shift the entire intellectual framework of content moderation from the enforcement of objective rules to the consideration of “context and narrative,” thereby allowing executives to engage in what amounts to viewpoint discrimination.

Private companies, of course, for the most part have the right to engage in viewpoint discrimination—something the government is prohibited from doing by the First Amendment. The problem is that when Twitter suspended Trump, it was operating less like a private company than like an extension of the federal government.

***

Among the most shocking revelations of the Twitter Files is the extent to which federal law enforcement and intelligence agencies came to view Twitter as a tool for censorship and narrative control. In part six of the Twitter Files, Taibbi chronicles the “constant and pervasive” contact between the FBI and Twitter after January 2020, “as if [Twitter] were a subsidiary.” In particular, the FBI and the Department of Homeland Security wanted Twitter to censor tweets and lock accounts it believed were engaged in “election misinformation,” and would regularly send the company content it had pre-flagged for moderation, essentially dragooning Twitter into what would otherwise be illegal government censorship. Taibbi calls it a “master-canine” relationship. When requests for censorship came in from the feds, Twitter obediently complied—even when the tweets in question were clearly jokes or posted on accounts with few followers.

Some Twitter executives were unsure what to make of this relationship. Policy Director Nick Pickles at one point asked how he should refer to the company’s cooperation with federal law enforcement and intelligence agencies, suggesting it be described in terms of “partnerships.” Time and again, federal agencies stressed the need for close collaboration with their “private sector partners,” using the alleged interference by Russia in the 2016 election as the pretext for a massive government surveillance and censorship regime operating from inside Twitter.

Requests for content moderation, which increasingly resembled demands, came not only from the FBI and DHS, but also from a tangled web of other federal agencies, contractors, and government-affiliated think tanks such as the Election Integrity Project at Stanford University. As Taibbi writes, the lines between government and its “partners” in this effort were “so blurred as to be meaningless.”

The Deputization of Twitter

After the 2016 election, both Twitter and Facebook faced pressure from Democrats and their media allies to root out Russian “election meddling” under the thoroughly debunked theory that a Moscow-based social media influence operation was responsible for Trump’s election victory. In reality, Russia’s supposed meddling amounted to a minuscule ad buy on Facebook and a handful of Twitter bots. But the truth was not acceptable to Democrats, the media, or the anti-Trump federal bureaucracy.

In 2017, Twitter came under tremendous pressure to “keep producing material” on Russian interference, and in response it created a Russia Task Force to hunt for accounts tied to Moscow’s Internet Research Agency. The task force did not find much. Out of some 2,700 accounts reviewed, only two came back as significant, and one of those was Russia Today, a state-backed news outlet. But in the face of bad press and threats from Democrats in Congress, Twitter executives decided to go along with the official narrative and pretend they had a Russia problem. To placate Washington and avoid costly new regulations, they pledged to “work with [members of Congress] on their desire to legislate.” When someone in Congress leaked the list of the 2,700 accounts Twitter’s task force had reviewed, the media exploded with stories suggesting that Twitter was swarming with Russian bots—and Twitter continued to go along.

After that, as described by Taibbi, “This cycle—threatened legislation wedded to scare headlines pushed by congressional/intel sources, followed by Twitter caving to [content] moderation asks—[came to] be formalized in partnerships with federal law enforcement.”

Late in 2017, Twitter quietly adopted a new policy. In public, it would say that all content moderation took place “at [Twitter’s] sole discretion.” But its internal guidance would stipulate censorship of anything “identified by the U.S. intelligence community as a state-sponsored entity conducting cyber-operations.” Thus Twitter increasingly allowed the intelligence community, the State Department, and a dizzying array of federal and state agencies to submit content moderation requests through the FBI, which Chan suggested could function as “the belly button of the [U.S. government].” These requests would grow and intensify during the Covid pandemic and in the run-up to the 2020 election.

By 2020, there was a torrent of demands for censorship, sometimes with no explanation—just an Excel spreadsheet with a list of accounts to be banned. These demands poured in from FBI offices all over the country, overwhelming Twitter staff. Eventually the government would pay Twitter $3.4 million in compensation. It was a pittance considering the work Twitter did at the government’s behest, but the payment illustrated a stark reality: Twitter, a leading gatekeeper of the digital public square and arguably the most powerful social media platform in the world, had become a subcontractor for the U.S. intelligence community.

***

The Twitter Files have revealed or confirmed three important truths about social media and the deep state.

First, the entire concept of “content moderation” is a euphemism for censorship by social media companies that falsely claim to be neutral and unbiased. To the extent they exercise a virtual monopoly on public discourse in the digital era, we should stop thinking of them as private companies that can “do whatever they want,” as libertarians are fond of saying. The companies’ content moderation policies are at best a flimsy justification for banning or blocking whatever their executives do not like. At worst, they provide cover for a policy of pervasive government censorship.

Second, Twitter was taking marching orders from a deep state security apparatus that was created to fight terrorists, not to censor or manipulate public discourse. To the extent that the deep state is using social media companies like Twitter and Facebook to subvert the First Amendment and run information psy-ops on the American public, these companies have become malevolent government actors. As a policy matter, the hands-off, laissez-faire regulatory approach we have taken to them should come to an immediate end.

Third, the administrative state has metastasized into a destructive deep state that threatens to bring about the collapse of America’s constitutional system within our lifetimes. Emblematic of the threat is the fact that “the intelligence community” has proven itself incapable of not interfering in American elections. The FBI in particular has directly meddled in the last two presidential elections to a degree that should call into question its continued existence. Indeed, the FBI’s post-9/11 transformation from a law enforcement agency to a counter-terrorism and intelligence-gathering agency with seemingly limitless remit has been a disaster for civil liberties and the First Amendment. We need either to impose radical reforms or scrap it entirely and start over.

The late great political scientist Angelo Codevilla argued that our response to 9/11 was completely wrong. Instead of erecting a sprawling security and surveillance apparatus to detect and disrupt potential terrorist plots, we should have issued an ultimatum to the regimes that were harboring Al Qaeda: you make war on these terrorists and bring them to justice or we will make war on you. The reason not to do what we did, Codevilla argued, is that a security and surveillance apparatus powerful and pervasive enough to do what we wanted it to do was incompatible with a free society. It might defeat the terrorists, but it would eventually be turned on the American people.

The Twitter Files leave little doubt that Codevilla’s prediction has come to pass. The question we face now is whether the American people and their elected representatives will fight back. The fate of the republic rests on the answer.