Elon Musk’s Grok AI violated Canada’s privacy laws and, at one point, was generating more than 6,000 sexual deepfakes every hour, according to Canada’s privacy commissioner.

Russia threw the kitchen sink at Hungary’s election: bots, deepfakes, “Maidan” paranoia, anti-Ukraine hysteria, and the full Kremlin-friendly propaganda machine. GRU, SVR, the Social Design Agency - all tried to interfere. Pro-Fidesz propaganda amplified anti-Ukraine hysteria. And still failed. The lesson is: Russia is dangerous - but not omnipotent. The myth of an all-powerful Kremlin is itself Russian disinformation. My comments to @Telexhu @stuthnagyniki #authoritarianinflation telex.hu/english/2026/06/12/…

DHS documented a systematic framework for tracking Russian disinformation in 2005. Two decades later, that baseline analysis reads like a user manual for what we're seeing unfold in real time. The framework identified how adversaries exploit social divisions and target information gaps to crack public confidence in democratic institutions. Putin's operatives have turned this into an industrial operation. What started as crude bot farms and obvious fake accounts has evolved into something far more dangerous: native-looking content that amplifies existing American voices rather than replacing them. The 2005 assessment caught Russian operations at their most primitive stage. Think spray-and-pray tactics across early social media platforms. Compare that to 2024, where the same adversaries have learned to weaponize authentic American influencers, hijack genuine political movements, and seed narratives that Americans then spread organically. They don't need armies of fake accounts when they can turn real Americans into unwitting distributors. The trajectory is clear. Russian information operations have moved from imitation to amplification. They've stopped trying to sound American and started making Americans sound Russian. The next evolution is already happening: AI-generated content sophisticated enough to fool detection systems, deepfakes targeted at local elections where verification resources are thin, and micro-targeted disinformation campaigns that exploit hyperlocal grievances. DHS built their 2005 framework around the assumption that foreign interference would remain foreign in origin and obviously artificial in execution. That model is dead. Modern Russian operations succeed precisely because they don't look foreign anymore. The vulnerability has shifted from technological to psychological. Early Russian bots were easy to spot because they couldn't replicate authentic American discourse patterns. Now they don't need to. They've mapped our political fault lines so precisely they can predict which authentic Americans will amplify which messages. They've turned our own polarization into their delivery system. Federal defenders are still playing catch-up to this evolution. The DHS framework treats foreign interference as an outside force penetrating American information space. But when Russian narratives get picked up by sitting members of Congress and broadcast on major networks, the traditional inside-outside distinction collapses. How do you counter foreign influence when it's being spread by domestic political leaders who may not even know they're doing it? The operational environment has fundamentally changed since 2005. Russian disinformation campaigns now launch simultaneously across Telegram, Truth Social, mainstream conservative media, and progressive activist networks. The same false narrative gets customized for different audiences and spread through their trusted information sources. By the time fact-checkers identify the original foreign source, millions of Americans have already encountered the story through channels they trust. Intelligence agencies know what's coming next. Russian operators are testing AI systems that can generate location-specific disinformation for congressional districts, create synthetic local news sources that build credibility over months before deploying false stories, and produce audio deepfakes of local officials discussing nonexistent scandals. The 2026 midterms will be the testing ground. State and local officials should expect Russian influence operations to go hyperlocal. Instead of trying to swing presidential elections, they'll target mayoral races, school board elections, and ballot initiatives where a few hundred votes can flip outcomes. Small-town newspapers won't have the resources to fact-check synthetic scandals about local candidates. Election officials in counties with populations smaller than a Moscow apartment building will face disinformation campaigns designed by intelligence professionals. The federal response remains trapped in the 2005 mindset that treats this as a cybersecurity problem. But you can't patch human psychology. Russian success doesn't depend on hacking voting machines or penetrating government networks anymore. They've hacked something more valuable: the American information ecosystem itself. Congressional oversight committees should focus on why U.S. intelligence agencies are still analyzing foreign interference through frameworks designed for simpler threats. The Russians documented in that 2005 assessment have spent twenty years studying American political psychology, mapping information networks, and testing influence techniques. Federal defenders need tools designed for adversaries who understand American politics better than most Americans do. Voters should assume any political content that confirms their existing beliefs and makes them angry is worth double-checking. Russian influence operations work by making Americans more extreme versions of themselves. They don't try to change minds anymore. They try to radicalize people who already agree with each other. The pattern is acceleration. Each election cycle, Russian operations get more sophisticated, more targeted, and harder to distinguish from authentic American political discourse. The 2005 baseline shows how far we've traveled toward a reality where foreign interference and domestic polarization have become functionally indistinguishable. Election officials preparing for 2026 should plan for disinformation campaigns that exploit local issues they've never heard of, delivered through trusted community voices, designed to suppress turnout in specific precincts. The Russians aren't coming to steal votes. They're coming to convince Americans that voting doesn't matter because the system is rigged anyway. That's the real trajectory. Not foreign control of American elections, but foreign amplification of American distrust until the elections don't matter because nobody believes the results. The 2005 framework documented the beginning of that process. We're watching the endgame. foreigninterference.org/post… #foreigninterference #DisinformationCampaigns

Oh god I just realized someone out there is probably making Boxxy deepfakes

Besides directly feeding the dehumanizing culture that creates malicious deepfakes or sexualized edits. If youre this ignorant you clearly dont care about her, only the fantasy in your head

🔴 #TurnoEnVivo | Josefa Barraza (@masterjosefa) y Camila Leyton, abogada sobre los deepfakes íntimos: “Hace un par de meses, a mí me están llegando mensajes anónimos diciendo que hay videos míos hechos con inteligencia artificial teniendo relaciones sexuales". Sigue la transmisión en vivo aquí: n9.cl/qsbum