Europe's Free Speech Failure and the "Censorship Industrial Complex" with Sarah B. Rogers
These notes were generated by AI and may contain inaccuracies.
David and the host are staying in a 300-year-old house where they've both hit their heads on the beams. This is both their first Davos after 24 hours. They're staying far away, as if not wanted close. Sarah Rogers finally got an invitation that didn't get lost in the mail, referencing an inside joke from the pod. At USA House, thanks to sponsors. Sarah Rogers is Under Secretary for Public Diplomacy at the State Department.
Friends and family asked "what is that?" Traditional diplomacy is between governments—ambassadors shake hands, make deals, solve wars. Public diplomacy addresses relationships between American government and foreign publics. Grown important with internet rise and Biden administration concerns about disinformation and malign influences on public views of America. Portfolio includes interacting with internet and information ecosystem, plus soft power like educational, cultural, sports diplomacy: World Cup, LA Olympics, Fulbright program.
Sarah focused on freedom of speech tensions between US standards and companies. US tech shifted back to stronger free speech after a period of moderation. Europe doesn't share this love.
First official trip to Europe coincided with large fine on X under Digital Services Act (DSA). US has stronger free speech approach since Founding and First Amendment than most West. Internet makes communication transnational, so European technocratic regulations bump against US commitments.
UK Online Safety Act (OSA)
Imposes age gating on broad upsetting content, requires risk assessments and removal of illegal content. UK bans major categories illegal there but not in US (where platforms, users, executives, primary regulator are). Active litigation against American websites not targeting UK—exist on US soil, host US users, discuss US politics—but fined for offending UK law.
EU Digital Services Act (DSA)
Requires EU states adopt minimum hate speech prohibitions, vaguer than US standards. US First Amendment requires clear prohibitions to avoid chilling effect on speech, especially on risk-averse corporations—leads to jokes becoming illegal. Also regulates transparency, competition, digital commerce, social media. Europe favors technocratic regulation over US rugged individualism. US position: Europe can build own platforms with own standards; these are US platforms with US standards, don't fine our users/platforms.
American companies abide by local laws abroad, but US issued sanctions on former EU official who threatened Elon Musk over X hosting Donald Trump interview in US Twitter Space—not for specific violative content, but act of hosting offended EU preferences. Offends US interests/values. US has jurisdictional concepts: website in California not regulated by Texas merely for existence. Europeans have laws in Europe; US has own laws, celebrating 250 years independence. Want interoperable online markets without giving up US free speech.
Government afraid of people criticizing it, protects power. Over 12,000 Brits arrested for speech acts in 2023 alone—more than in Russia, China, Turkey. Not just under OSA/DSA; portals applying existing laws like Communications Act, laws against inciting racial/religious hatred (different from US incitement standards).
Examples of UK Arrests
- Comedian Graeme Lahan tweeted woman seeing penis in ladies' room should kick guy in balls—dragged from airport, devices confiscated, jailed overnight, lost heart medication. Offended provocative speech law.
- Footballer Joey Barton called someone "bike nonce" (insinuating effeminate for bike riding)—suspended prison sentence.
Public polls oppose this; Brits say unacceptable. Regulators defend as less totalitarian than China, but arresting 12,000/year for speech, dragging people from airports for offending transgender activism dogmas, risks becoming like China. Why US defend/pay for such allies in NATO/proxy wars vs Russia? Shared values questioned—free world cornerstone is freedom of speech, especially uncomfortable speech criticizing leaders.
Much about immigration. Judge gave 2-year sentence for criticizing UK's open immigration policies.
Lucy Connelly Case
31-month sentence to suburban mother after man Axel Rudabanka stabbed 7-, 8-, 9-year-old girls at birthday party, causing unrest. She tweeted anti-migration: "If this is what migration does to our country, burn down the migrant hotels for all I care" (paraphrase; deleted after feeling bad). Inflammatory but legal in US. Meanwhile, pedophiles/child sex offenders get minimal/no prison—leads to two-tier policing epithet: opponents of mass migration face different justice than Sharia agitators or child porn downloaders.
UK OSA provisions phasing in, including AI; active US litigation e.g., 4chan (no-holds-barred meme site, source of cat memes, Occupy Wall Street; bans only child porn). UK demands payment for not policing per UK speech laws.
€140 million fine on X during Sarah's Europe trip. X has political valence; Biden wanted ways to go after Elon post-acquisition. Not advocating for one company—if General Motors fined for First Amendment, US would respond. Others fined under EU digital regs. Infographic shows EU revenues from fining US tech exceed other metrics—suspicion of de facto tax on pretext.
DSA as Censorship Tariff
DSA like digital speed trap, disproportionately hits US companies like tariff. Europe can have tariffs, but changes US tariffs—part of trade negotiation. Cost of DSA censorship apparatus levied on large (mostly American) companies via intricate standards. Easier politically to tax Americans than Europeans.
Freedom of speech declining in Europe/UK, rising in Middle East e.g., Riyadh comedy festival (rules: no criticizing kingdom/religion, but go after own). South Korea requires social security number to post online. Censorship protects power from criticism; better to hear/adjust than silence feedback. EU/UK mass migration unpopular—listen/adjust instead.
South Park deranged attacking President Trump. Jimmy Kimmel/Charlie Kirk kerfuffle resolved quickly—no government censorship, just network affiliates upset over untrue/malicious remarks. Brendan Carr rethought. Trump/admin has thick skin. Disturbing closest allies (Western culture/history) moving to more censorship; glad State Dept under Trump/Secretary Rubio pushing back via trade tools, visa denials, condemnation.
18 months ago deepfakes failed uncanny valley; now Grok images, Nano Banana (Google) fool quick glances. US/French protections for cartoonists mocking public figures—how different with AI where public can't tell?
At technological frontier. After Charlie Hebdo cartoonists murdered (Sarah represented on First Amendment; knew him), Americans wore "Je Suis Charlie" t-shirts. French stood up for free speech under threat from religious zealots. Different kind of religious zealotry to not allow
Reflecting on European consensus on free speech since the Charlie Hebdo episode is sobering. America takes pride in being a civilization where Charlie Kirk and Charlie Hebdo can both speak. Making fun of public officials and pointing out when the emperor has no clothes is essential in a democracy for self-governance.
The point of parody is that you can tell it's parody. Deep fakes depict public officials falsely in ways people can't tell as satirical or non-authentic, undermining the parody tension.
Whenever we reach a new technological frontier like deep fakes, there's temptation to enact a flurry of new regulations. History shows similar reactions to the printing press (thought to end the world), telegraph (worries about disinformation and attention span), and film strip (people thought the train was coming at them). The impulse to regulate rather than allow adaptation and give freedom the benefit of the doubt is vindicated over time.
America and Europe have strong legal remedies against defamation. If someone creates a false image of a public figure that reasonable viewers believe, you can sue for defamation. Child protection and underage laws are strong. Using AI to break the law, like cyber hacking, leads to prosecution for cyber crime. Existing laws apply to AI; use them before creating AI-specific ones. If they don't cover it, have a thoughtful discussion.
Edge cases like voice clones fall under fraud or wire fraud statutes. Regulatory tweaks could make fraud easier to detect, such as watermarking AI images to mitigate tort liability for providers. When inventing mass-scale capital markets, old fraud laws were supplemented with fine-tuned securities regulations like annual disclosures, without fundamentally changing treatment of false information.
Don't put technological innovation back in the bottle, especially with rivals like China developing AI aggressively. Cocooning in safety hurts standing in that race.
Before his appointment by President Trump and Secretary Rubio, the speaker was a litigator fighting cases; now as diplomat, optimism is sounded. Ordinary Europeans are uncomfortable with comedians dragged out of airports, just as they were with comedians murdered for offensive cartoons. Polls in Europe reflect this sentiment. Productive conversations have occurred, with hope for more.
If acute, is Europe prepared for American social networks blocked by IP address? Platforms don't need European money. Would European countries demand their own versions, like a UK-specific X?
One resolution to transnational issues is geofencing. UK enforcement like Ofcom has sued small American websites despite geofencing and blocking UK IPs, which is ridiculous. Users can use VPNs costing $30/year, as people in Iran do. Countries without Charlie Hebdo tradition like Russia or China admit censorship and block sites with the Great Firewall.
UK putting up a Great Firewall is technologically feasible but not politically; British people want to be free. If Keir Starmer announced it, he'd be out of office. They prefer subtle fines over obvious blocks. Pushback says no. Elon Musk is principled and can pay fines.
President Trump's election changed US free speech direction. Under Biden, cases like Biden v. Missouri, Twitter Files, and disclosures showed administration pressuring social networks for censorship. Without Trump, censorship would continue. Zuckerberg follows the wind for system growth; Elon is an outlier in standing up for free speech.
FBI pressured tech companies on tweet tones deemed damaging, suppressing COVID discussions like not needing vaccines for kids. Pretext was disinformation, distended to include true info promoting adverse narratives per NGO white papers.
Misinformation is false; disinformation is bad, polluting democracy if wrong candidate wins. Suppressed true info included: vaccines not completely preventing transmission (mitigate but not sterilizing); virus possibly leaking from lab (House committee and CIA assessments). Government nudged platforms: can't force under First Amendment but appreciate removal to stay in good graces.
Be vigilant about free speech; if not, it gets taken away.
NGOs instigate regulators with cases. Many are US-based, going to European regulators as an end-run around US First Amendment to censor US material.
Leaked emails from Center for Countering Digital Hate (British NGO, leader visa-sanctioned) with US Democrats and Keir Starmer allies: priority to kill Musk's Twitter (American company) to suppress US speech; second, instigate UK/EU regulatory action. Entity takes government money to get foreign governments after US businesses.
US NGOs worked with US government sending forceful (not technically coercive) emails to Twitter/Meta, replicating EU DSA to dodge First Amendment. DSA requires member states designate NGOs as trusted flaggers with privileged reporting channels on Twitter for hateful content, prioritized.
Under Biden, similar: government arranged first-trunch priority for NGO reports. Zoom meetings showed operatives saying couldn't do under First Amendment, but NGO would.
Firms pressure advertisers to cancel (started conservative, moved liberal: Howard Stern, Rush Limbaugh). More pernicious: pressure Cloudflare, Amazon, PayPal, Stripe to debank/demonetize; YouTube shadowbans/labels (e.g., All-In COVID scientist discussions).
Without monetization, scaling is impossible. One successful Supreme Court case on viewpoint-based debanking: NRA v. VUO (speaker's case, won). New York financial regulator told institutions better for enterprise risk management if no pro-gun business.
Finance has reputational risk for safety/solvency, expanded by ESG to disfavored speakers' bank accounts. Supreme Court ruled government can't adversely apply regulations to choke viewpoints via risk-averse middlemen like banks.
Debanking insidious: middlemen lack skin in game, follow in-house counsel fearing regulators. Common theme: government uses intermediaries (banks, dark NGOs funded by government posing as non-government for fact-checking, FBI on social networks) for dirty work when direct action violates First Amendment—nefarious and hard to detect.
Zuckerberg pragmatist: FBI calls before FTC merger make nice. Regulators with capricious cudgel over internet lead companies to impress even if not speech-based. Techno merger under Biden: showed how woke to regulators; now MAGA-lined.
Labeling seems transparent but depends on who/why. Disinformation NGOs sent lists to credit card/payment processors (government-funded, implicit imprimatur) suggesting not fund sites; sites/viewers unaware. Good labeling: X community notes—read tweet, see ranked community notes. Game changer over biased fact-checking.
Community notes catch fake images/articles quickly, notify users. Genius algorithm: promotes only if users who usually disagree reach consensus among rivals. Grok AI fact-checks well: ask "what's the truth?" Trending right direction.
No need for bureaucrats/politicians/regulators on truth—we have community notes, AI, users, defamation lawsuits.
Grateful for vigilance protecting First Amendment. [Applause]
Keep All-In Podcast in your library
Save the videos and channels worth coming back to, and find them again in one place.
