Europe's Free Speech Failure and the "Censorship Industrial Complex" with Sarah B. Rogers

Summary

• Rogers argues that Europe’s speech rules now threaten both American constitutional norms and American technology companies. The UK Online Safety Act and EU Digital Services Act turn local speech prohibitions into platform obligations, including against websites based on American soil that host American users and often discuss American political topics. Her line is jurisdictional, not absolutist: Europe may govern European services, but America will not “give up American freedom of speech in the bargain.”
• The UK’s reported 12,000-plus speech arrests in 2023 are Rogers’s evidence that the dispute is no longer theoretical. She cited comedian Graham Linehan’s airport arrest, footballer Joey Barton’s suspended sentence and Lucy Connolly’s 31-month sentence for an inflammatory anti-migration post. Sacks’s interpretation: censorship protects incumbents from feedback on unpopular mass-migration policies; Rogers warned that raising children under such rules means Britain “might not have a different culture than China for long.”
• The economic mechanism is a “censorship tariff” imposed disproportionately on large American platforms. Rogers recalled an X fine of roughly 140 million, though she was unsure whether the denomination was euros or dollars; Calacanis called the DSA a “censorship tariff,” while Rogers described it as a digital speed trap or effectively a tariff. That converts a speech dispute into trade policy: fines, reciprocal tariffs, visa restrictions and sanctions all become possible US responses.
• Europe’s attempt to regulate American speech without openly blocking American networks may prove politically unstable. Geofencing has reportedly failed to satisfy some UK enforcement efforts, yet Sacks said a British “Great Firewall” is technologically feasible only to a point and politically implausible because “British people want to be free.” Rogers’s diplomatic hedge was deliberate: “I can’t promise a panacea,” but public opposition to speech arrests gives her grounds for a “gong of optimism.”
• On AI deepfakes, Rogers and Sacks favor applying existing defamation, fraud, child-protection and cybercrime laws before creating a sweeping new regime. Watermarking or other tailored detection rules might help, but the guidance was “fine-tuned” intervention, not trying to put innovation “back in the bottle.” The strategic risk is that US “safetyism” constrains domestic AI while China develops aggressively.
• The discussion describes a censorship-industrial complex that routes government pressure through NGOs and foreign regulators. Rogers cited leaked emails involving the Center for Countering Digital Hate and politicians that allegedly prioritized “kill Musk’s Twitter” and instigating UK and EU action; the DSA’s “trusted flaggers” provide a privileged reporting lane. During COVID, “disinformation” was stretched to include truthful material carrying an “adverse narrative,” including claims that vaccines did not completely stop transmission and that a lab leak might have occurred.
• The recurring enforcement pattern is pressure on a “risk-averse middleman” rather than a direct speech ban. Rogers connected platform moderation, debanking, payment processing, advertising and discretionary merger approvals to the same leverage model, citing her Supreme Court victory in NRA v. Vulo. She prefers visible, contestable correction systems such as Community Notes—whose “consensus amongst rivals” design leaves the original post readable—over government-backed blacklists quietly circulated to banks and processors.

Deep dive

1. Public diplomacy now runs through privately owned information systems

• Rogers defined her post as managing the relationship between the US government and foreign publics, rather than traditional government-to-government diplomacy. That portfolio now includes the internet’s information ecosystem alongside educational, cultural and sports programs such as Fulbright, the World Cup and the Los Angeles Olympics.

• Her foundational claim is that America’s First Amendment tradition has always been more speech-protective than most of the West, but transnational platforms now force those legal cultures into collision. American technology companies briefly moderated in line with more censorious domestic norms; as the US and those companies shifted back, European regulation became the sharper point of conflict.

• The jurisdictional principle is narrower than “American companies need not obey foreign law.” Rogers accepted that companies operating abroad follow local rules, but rejected foreign punishment based merely on an American website’s existence or American political speech. Calacanis distilled the position: Europe can build platforms with European standards, while America protects “our platforms” and “our standard.”

2. Britain’s speech arrests turn regulatory theory into prison time

• Rogers said slightly more than 12,000 Britons were arrested for speech acts in 2023—more, she said, than in Russia, China or Turkey that year. Regulators answer that Britain remains less chilling, so more people risk offending; her rebuttal was that normalizing an airport arrest and device seizure may erase that cultural distinction over time.

• Her sharpest example was Graham Linehan, arrested at an airport after joking that a woman confronting a penis in a women’s restroom should kick the person responsible. Rogers said he was detained overnight, lost his devices and, to her recollection, lost access to heart medication—conduct nowhere near the American incitement standard.

• Migration policy supplied the most consequential disagreement between citizens and government. Rogers cited Lucy Connolly’s 31-month sentence for saying, after what she recalled as the stabbing of three girls—aged 7, 8 and 9—at a birthday party, “burn down the migrant hotels for all I care”; Connolly deleted the inflammatory post after regretting it, but still received prison time. UK activists call the perceived disparity with treatment of other crimes, including child-sex offenses, “two-tier policing.”

• Sacks broadened the issue from civil liberty to alliance logic. In the NATO context, he said, Americans hear about shared history and shared values, but “the cornerstone of a free world” must be protecting uncomfortable criticism, not merely agreeable expression.

3. Platform penalties are becoming trade instruments

• Rogers described the Online Safety Act as a portal applying Britain’s many existing speech offenses to the internet. Ofcom has pursued US websites that neither target Britain nor sell there; even geofencing UK addresses reportedly has not ended one case. Jason noted that British users could still evade the block with a VPN.

• The EU’s DSA combines content rules with transparency and competition regulation, requiring member states to maintain at least a floor of hate-speech prohibition. Rogers’s legal objection is vagueness: broad rules imposed on “large risk-averse corporations” create a chilling effect, just as ambiguous workplace restrictions eventually make ordinary jokes effectively forbidden.

• The commercial implication is Rogers’s “censorship tariff”: large platforms, predominantly American, bear the cost of a detailed compliance apparatus and potentially enormous fines. She recalled X receiving a penalty around 140 million—uncertain whether euros or dollars—while Calacanis called the DSA a censorship tariff and Rogers described it as a digital speed trap that could function as a tariff even if Europe formally classifies companies by size rather than nationality.

• The conversation pointed to trade measures, including reciprocal tariffs, visa denials and sanctions as possible US responses. Rogers cited a sanctioned former EU official who threatened enforcement when X planned a US-hosted interview with Donald Trump—not over any identified violation, but over the possibility that hosting the interview itself might offend European preferences.

4. Europe cannot easily admit what full enforcement requires

• Calacanis asked whether an irreconcilable divide ends with country-specific networks or American platforms abandoning Europe. Rogers identified geofencing as one technical compromise, but the reported UK insistence that even blocked American sites remain accountable makes that settlement doubtful.

• Russia and China openly block services they reject; the hosts argued that UK authorities prefer fines to acknowledging an equivalent choice. Sacks called a British firewall technologically feasible “to a point” but politically implausible: if Keir Starmer openly barred American social media, he believes “he’d be out of office.”

• Rogers’s outlook remained expressly uncertain. “I don’t know where things are going” and “I can’t promise a panacea,” she said, but polling and her private conversations suggest ordinary Europeans dislike comedians being arrested just as they opposed cartoonists being murdered. That warranted a deliberately diplomatic “gong of optimism.”

5. Existing law should get the first shot at AI harms

• Deepfakes complicate parody because parody normally signals its own falsity; an image believable as authentic may instead be defamatory. Rogers nevertheless defended ridicule of public officials as democracy’s essential “emperor has no clothes” function and invoked a society where both Charlie Kirk and Charlie Hebdo can speak.

• Her historical analogy: printing presses, telegraphs and film each produced panic about destabilization, falsehood or human perception, while societies eventually adapted. “Give freedom the benefit of the doubt” before enacting a flurry of rules, because regulatory restraint at technological frontiers tends to be vindicated over time.

• Calacanis called AI-assisted voice cloning “just fraud,” and Sacks suggested wire fraud; Rogers said AI-assisted hacking remains a cybercrime. Sacks suggested “fine-tuned” additions such as watermarking or other detection tweaks, analogous to specialized securities reporting built atop general fraud law, but warned that cocooning the US in “safetyism” would weaken it against China’s aggressive AI development.

6. The censorship system works through deniable intermediaries

• Rogers said the answer to whether activists use Europe as an end run around the First Amendment is already “yes.” She cited leaked emails involving NGOs including the Center for Countering Digital Hate, Democratic politicians and figures close to Keir Starmer that allegedly named “kill Musk’s Twitter” as the first priority and UK/EU regulatory action as the second; the NGO’s leader was a target of some US visa sanctions.

• The DSA formalizes NGO “trusted flaggers,” whose complaints receive privileged platform review. Rogers compared that structure with US agencies arranging priority channels during the Biden administration, including recorded meetings where government operatives acknowledged, “We couldn’t do this under the First Amendment,” but an NGO participating in the call could act instead.

• David argued that the term “disinformation” had been stretched beyond false claims to cover truthful information that promoted an “adverse narrative.” He said suppressed COVID claims included that vaccination did not completely prevent transmission, which he said turned out to be true, and that the virus might have leaked from a laboratory. He also said the lab-leak assessment was shared by a House committee and the CIA and was viewed by the government as more likely than not.

• Sacks summarized the common architecture: government pressures an NGO, bank, processor or platform to do what it cannot constitutionally do itself. Rogers’s NRA v. Vulo victory established that a regulator cannot weaponize “enterprise risk management” or reputational-risk rules to choke off disfavored viewpoints through financial institutions.

7. Transparent correction beats covert blacklisting

• Rogers extended the intermediary problem to discretionary merger approval: if a regulator possesses a “capricious cudgel,” companies advertise ideological alignment to please whoever holds it. She thought that incentive may have shaped the Jimmy Kimmel episode; Sacks still maintained there was “no real censorship,” noted Kimmel returned within two or three nights and said Brendan Carr “probably shouldn’t have said what he said.”

• Government-funded disinformation groups did not merely publish assessments, Rogers said; they circulated lists to credit-card companies and payment processors, quietly pressuring outlets’ ability to monetize. The publisher and audience might never know why funding disappeared, while the government’s implicit imprimatur magnified the pressure.

• Her preferred counterexample is Community Notes: the disputed post stays visible, the proposed correction is inspectable and consensus is sought among users with rival views. Sacks contrasted that feedback loop, including notifications to users who shared false material, with newspaper corrections buried where the original audience may never see them.

• Grok adds another checking layer. Rogers said it is not perfect, while Sacks said it is “trending in the right direction.” Their conclusion was not that error disappears, but that users, Community Notes, AI and defamation suits offer visible remedies without installing bureaucrats as final arbiters of truth.