Who was president in 2020?
Hello! It’s been a while since I’ve emailed – work has been busy. I just turned in the epilogue for the upcoming paperback version of Invisible Rulers, an academic paper that took three years to get out the door is finally finished, and I published a whitepaper with a colleague at Carnegie about the need for institutions to fundamentally rethink comms strategies in the age of modern influence dynamics.
But in this newsletter I want to take on something simultaneously more basic and yet seemingly insurmountably complex: who was president in 2020?
Yesterday I watched Senator Eric Schmitt (R-MO) run a Senate Commerce Committee hearing, spending hours promoting claims about a "Biden censorship regime" that had managed to censor people, such as witness Sean Davis of The Federalist, during the 2020 Presidential campaign. Last week, Senator Ted Cruz (R-TX) – on whose behalf Schmitt was running the hearing – had released a report making an even bolder claim: that this censorship regime had begun its work in 2018.
This would be darkly funny if the consequences weren’t so serious. Because this past month has given us one of the clearest contrasts we've had in a while about what censorship actually looks like in the United States – and how completely the narrative machinery has disconnected from reality.
In mid-September, FCC Commissioner Brendan Carr issued barely-veiled threats to ABC over Jimmy Kimmel's comments in the wake of the assassination of Charlie Kirk, saying the network could handle the situation "the easy way or the hard way." Within hours, Kimmel was suspended.
Later that month, Google sent a letter to Rep. Jim Jordan's House Judiciary Committee acknowledging that during the Biden administration, government officials had urged content moderation related to COVID-19. But Google also made clear, repeatedly, that it had resisted those requests and maintained its own independent policies.
One incident was jawboning backed by regulatory threat. The other described pressure met with pushback.
The Kimmel affair did spark significant outcry…from civil liberties organizations like FIRE and, notably, from several prominent voices on the right – including Ted Cruz. But that outrage, while genuine, operated in a different register than the sustained, institutional narrative Cruz and Schmitt were promoting yesterday. That narrative has been years in the making, reinforced through congressional investigations, litigation, and a dedicated media ecosystem that continues to treat debunked claims as established fact.
The 2018 claim isn't a mistake. It's the whole game.
So let's look at what actually happened when these claims were investigated.
A "censorship regime" that disappears under scrutiny
The myth of the “Biden censorship regime” – aka the "censorship-industrial complex" – didn't emerge fully formed. It began as a familiar tactic: working the refs to get content moderation policies shifted in a direction favorable to one’s political side. This is something that inevitably happens on centralized social media platforms; politicians across the spectrum do it, everywhere in the world. Liberal politicians often allege harms from particular types of content; conservatives have spent the better part of a decade alleging anti-conservative bias. Many of these allegations aren’t borne out by studies or investigations, but they generate enough public outrage that platforms are sensitive to the accusations. There’s often a grain of truth – content moderation is impossible to do well at scale (Masnick’s Impossibility Theorem) so there are ample anecdotes out there to motivate grievances.
But since the House flipped in 2022, right-wing ref-working evolved into something more sophisticated: actual lawfare. Claims of intentional viewpoint-based discrimination moved from op-eds and social media influencer complaints into congressional hearings that operated in tandem with vexatious lawsuit mills. Blatantly misleading lies showed up in judicial opinions obtained via forum-shopped litigation, and in official Congressional reports via the Weaponization Committee, where they took on a veneer of legal and empirical legitimacy. The allegations expanded from mere bias to a conspiracy theory that the government had violated the First Amendment through coercion and the supposed conscription of dozens of NGOs as “cut-outs” into an elaborate cabal that “suppressed entire narratives” as part of an attempt to steal the 2020 election from Trump. Half-cocked analyses and selectively-cropped emails from The Twitter Files were the primary evidence offered to back the allegations up.
Since we are still holding congressional hearings on this theory as recently as yesterday, I laid out the evidence gaps in a Lawfare article covering the three recurring claims: the Murthy v Missouri case that Schmitt brought when he was AG of Missouri; the Twitter Files; and the investigation of the Weaponization Committee.
The distinction between pressure and coercion isn't just semantic – it's constitutional. First Amendment law draws a bright line between a government asking and a government threatening. The courts have consistently recognized that communication between government and platforms is not, in itself, censorship. And that’s why it’s important to understand reality vs allegation.
Three investigations, three collapses
The mere filing of the Murthy case serves as evidence of the censorship regime for MAGA politicians – particularly Senator Schmitt, who kept pointing to it during yesterday’s hearing. The case involved a forum-shopped district court judge who fabricated quotes and misconstrued evidence; a Fifth Circuit Court of Appeals decision that overturned 9 of 10 provisions in the judge’s original injunction; and then SCOTUS tossed the case for standing. It was not a win! Justice Amy Coney Barrett's opinion was meticulous in its examination of the record. She wrote that the factual findings by the district court had been "clearly erroneous”! She showed that platforms had "independent incentives to moderate content and often exercised their own judgment," with many policies predating any government communications.
Most devastatingly for Schmitt’s case, Barrett noted "the lack of specific causation findings with respect to any discrete instance of content moderation." The plaintiffs couldn't show that government pressure actually caused the takedowns they claimed as injuries! Remarkably, two of the plaintiffs – physicians Jay Bhattacharya and Martin Kulldorff, who built significant public profiles as wrongly censored "COVID contrarians" (Jay is now the head of the NIH) – didn't appear to have been mentioned by any government defendants at all. Without that causal link – without evidence that platforms removed content they otherwise would have left up because of government threats – there is no First Amendment violation.
In other words, the record in Murthy showed platforms making independent editorial decisions in accordance with their policies. That's not censorship. That's editorial judgment – in fact, it’s platforms exercising their First Amendment rights. Senator Schmitt would rather people not understand this.
The Twitter Files can’t find the 22 million tweets
The "censorship regime" narrative points not only to Murthy, but to two other pseudo-investigative efforts: the Twitter Files, and Rep. Jordan's Weaponization Committee. Each is offered as definitive proof of government coercion. And yet each, when examined, undermines the claim.
In March 2023, Matt Taibbi and Michael Shellenberger testified before Congress with explosive allegations: 22 million tweets had been censored by an academic cabal working with platforms at the direction of federal agencies during the 2020 election (who was President?). They claimed ‘secret CIA agents’ were behind the operation. They claimed Stanford demanded platforms censor "true stories of vaccine side effects." For months afterwards they wrote about a vast web of NGOs who’d supposedly colluded with the government.
After those targeted produced emails and work product under subpoena, every major allegation collapsed:
- The "22 million censored tweets" became 4,700 URLs flagged overwhelmingly by undergraduate student researchers across all platforms in 2020, focused on rumors related to voting and allegations the election was rigged, which appeared to violate platform policies.
- Platforms ignored 65% of those URLs entirely; the remaining 35% (~1,645) were overwhelmingly labeled, not removed. The handful of tweets that Jordan highlighted in his report were not identified by any deep state actor (since that hadn’t happened) but by students – and were still up.
- The lie about some demand to ‘censor vaccine side effect’ tweets came from Taibbi and his research assistant Andrew Lowenthal literally cutting an email in half and misrepresenting a bullet point in a list of content categories the project described studying – with no demand made anywhere therein.
- There were no federal agencies directing or funding the work or laundering requests. There were no secret portals.
Jordan's committee considered all this evidence. Their final report, unable to prove mass censorship, moved the goalposts and resorted to complaining about Stanford's document production timeline and characterizing Jira project management software as a "censorship arsenal."
Jordan's own witnesses contradict his narrative
But perhaps the most damaging evidence against the "censorship regime" comes from Jordan's committee’s interviews. Over a two-year span from 2023 to 2024, the committee conducted interviews with dozens of staff from Google, Meta, and other platforms, zeroing in on whether the government had exerted improper pressure.
The transcripts – all 12,000 pages, released in a document dump right before Christmas 2024 – show executives telling the same story as what Google wrote in their recent letter: despite outreach from government officials, platforms retained their independence.
When pressed about the “months of pressure” Jordan alleged influenced YouTube policy, a former public policy manager responded: "This is not an accurate characterization… YouTube doesn't change policies… as a result of pressure from the White House."
Asked if an email from White House official Rob Flaherty stating “we’d love to get into the habit” of regular communication was perceived as threatening or coercive, the manager replied: "I did not take that as a threat… I did not take that as coercion."
This same theme appeared repeatedly across dozens of interviews across tech companies. Even in instances where White House emails or meetings were mentioned, the strongest language used by witnesses to describe the tone was that officials had shown “frustration” – but they consistently stopped short of labeling it coercion or threats.
All of this testimony was known to Jordan before he heralded Google’s letter as proof of a censorship regime. He just chose not to feature it in any of his reports.
Why the evidence doesn't matter
Here's what makes this particularly insidious: none of these evidentiary collapses have mattered for the narrative. The original claims – 22 million tweets, secret portals, deep state agents – continue to circulate in the media ecosystems where they first took hold. Legal filings still reference them. Senator Schmitt continued to air them from the chair in yesterday’s hearing.
The audience for the sensational claim is not the same audience for the correction. And in a fractured media environment, you can keep a narrative alive indefinitely simply by never acknowledging the evidence that undermines it.
The politics of selective memory
I spent a few days last week at Vanderbilt’s excellent annual Global Free Speech Summit (I attended last year as well). I gave a talk on middleware; my longtime personal feeling is that ref-working is best managed by giving users more direct control of their feeds, though that’s a subject for a different newsletter. I also participated in a live recorded episode of The Good Fight Club podcast with Yascha Mounk, Jonathan Rauch, and Jacob Mchangama, who I always enjoy talking to and debating even (especially!) when we disagree. In one rather funny exchange, Yascha asked me about content moderation and whether we are all free speech hypocrites…a worthy thing to debate, though he raised the example of the moderation of Hunter Biden’s laptop as if it had happened under the Biden Administration.
Who was president in 2020?
The reframing of platform content moderation as synonymous with censorship, and as something that only one “side” wants, has been an incredibly effective campaign. For the record, I thought the Hunter Biden call by Twitter was a very bad call. We didn’t weigh in on it at SIO because it was out of scope for our work, but I said so at the time conversationally on Twitter, in podcasts, in my book, etc. And yet: it was a private company’s bad call. There was no “Biden regime” silencing the story to protect itself, and on platforms where the story was not blocked (for the ~48 hours Twitter did so) it was shared hundreds of thousands of times. This is one of the canonical examples of the Streisand Effect! And yet, someone came up to me after the panel to tell me that if the FBI hadn’t demanded the censorship of the Hunter Biden laptop, Donald Trump would have won.
This kind of myth-making isn't just a bug in the current discourse – it's a core feature. When you trust Jim Jordan, when his framing reaches you through a curated media ecosystem that amplifies his perspective while filtering out contrary evidence, you're going to believe the Biden regime “got to” Twitter somehow through rogue FBI agents. You're not going to see the facts that emerged from thousands of pages of testimony or from Justice Amy Coney Barrett's meticulous Supreme Court opinion finding "clearly erroneous" factual claims and zero causal links between government requests and platform actions.
When platforms stop saying no
YouTube just settled a $24.5 million lawsuit with Trump over his post-January 6th deplatforming – a case it was likely to win and that had already been administratively closed. Google committed in its letter to Jordan to never using "so-called fact-checkers" (which Jordan considers “censors”) – a policy it wasn't employing anyway, but now has publicly forsworn, ironically, under congressional pressure.
ABC suspended Kimmel hours after an FCC commissioner's threat – a regulatory official wielding government power to silence political speech.
The companies that rightly resisted Biden-era requests seem to be folding under Trump-era threats. That's the difference between pressure and coercion, between persuasion and power. As Jonathan Rauch put it during our podcast discussion, "This is a different kind of thing. This is not just turnabout hypocrisy.”
The danger isn't just hypocrisy, though there's plenty of that. It's that we're watching the construction of a permission structure for actual government censorship, built on the scaffolding of a manufactured scandal.
If you can convince people that Biden ran a secret regime to silence conservatives – in 2020! – then you can justify doing precisely that from a position of actual power, as a form of retribution. If you can get the public to accept that platforms were captured by the left, then there's no contradiction in capturing them for the right. The narrative justifies the retaliation.
I support anti-jawboning legislation; I think what FIRE has broached, for example, has real merit, and that government content moderation requests should be made transparently – all of government, Congress included. Maybe having more facts and evidence out there in the world can help…though I’m not optimistic, because ultimately people need to hear them from voices they trust, and there has to be incentive for those voices to tell accurate stories.
But in the meantime, the Big Tech platforms that said no to prior administrations can still say no to Trump. The question is whether they will – or whether actual coercion is becoming normalized. We're about to find out which they value more: editorial independence or regulatory peace.