Risks of Chinese and Russian Media Influence

This paper argues that China and Russia have used AI in foreign information manipulation since at least 2022. AI helps them produce more content, faster, cheaper, in more languages, and with more personalization. China is ahead. It has stronger national AI capabilities, domestic tools, large datasets, and a state-controlled tech ecosystem. Chinese-linked actors use AI for propaganda, fake personas, translated content, deepfakes, sentiment monitoring, data mining, and targeting individual users online. Russia has high ambitions but weaker AI capabilities. Sanctions, limited resources, weak hardware access, and talent constraints slow it down. Still, Russia uses open-market AI tools effectively to scale old tactics, fake news sites, fake journalists, cloned voices, deepfakes, and automated propaganda networks. The key difference is this: China uses AI more strategically, with better infrastructure and more precise targeting. Russia uses AI more opportunistically, to create confusion, volume, and distrust. The paper warns that future risks include agentic AI, more automated manipulation, harder attribution, more personalized disinformation, and growing dependence on Chinese AI infrastructure. It recommends combining technical AI-based countermeasures with traditional democratic resilience tools, regulation, platform accountability, media literacy, and stronger detection capacity.

READ ME: The 4th EEAS Report on FIMI just dropped. The European External Action Service (EEAS) released its most comprehensive mapping of Foreign Information Manipulation and Interference (FIMI) activity in 2025; it's filled with evidence-based analytics and infographics, covering 540 documented global incidents across 2025. A few things that stood out to me: The use of AI tools in detected FIMI incidents jumped ~259% compared to 2024, with synthetic audio, video, and text now embedded as standard tools across both Russian and Chinese operations. We're seeing AI used to create the "Firehose of Falsehood" and to enable "Spamouflage": Flood the zone, erode trust, repeat. Election interference continues to be a primary focus of Russian FIMI activity, and it follows a consistent three-phase logic: delegitimize leadership months before a vote, weaponize domestic divisions as election day nears, then undermine electoral integrity at the final stage. We've seen it in Germany, Moldova, Romania, and the same infrastructure is already warming up for Armenia's June 2026 elections. FIMI is a supply chain problem, not just a content problem. The report's new Deterrence Playbook reframes the challenge: by identifying and targeting the financial, technical, and organizational enablers behind FIMI operations, it becomes possible to raise costs and reduce the long-term viability of influence campaigns, rather than playing whack-a-mole with individual incidents. Full report linked below. Worth your time. 👇

OpenAI's latest threat intelligence report exposes an unsettling reality: advanced language models are being systematically integrated into the arsenals of cybercriminals, authoritarian regimes, and disinformation operators. Analysis of over 40 disrupted networks reveals methodologies that should concern every cybersecurity and legal professional. The Key Finding: AI as Optimizer, Not Revolutionary Contrary to speculation about entirely new threat vectors, criminals are taking a pragmatic approach - embedding ChatGPT into existing operational workflows. Russian-speaking operators leveraged the model for malware tooling development, Korean-language groups for phishing script creation, and Cambodian-based scammers for error elimination in fraudulent campaigns. AI serves as proofreader, translator, and code generator within proven criminal frameworks. The report documents two phenomena of particular significance for threat analysts. First, adversaries are already employing multi-model strategies: using ChatGPT to generate prompts for other AI systems, creating complex tool chains. Second, we're witnessing the emergence of recidivist operations - disinformation campaigns, likely of Russian origin (the "Stop News" network), immediately returning with fresh accounts after disruption. The Chinese Dimension Particularly alarming are cases of AI exploitation by entities linked to the People's Republic of China. OpenAI identified "Operation Nine-emdash Line" a sophisticated covert influence campaign leveraging the model to generate content supporting Beijing's narratives. These are no longer experimental probes but systematic information operations conducted by a state with unlimited resources. What makes this report particularly valuable is its candor about operational realities. Threat actors aren't inventing new attack categories; they're making existing ones faster, cheaper, and harder to detect. The Russian-speaking malware developers weren't building revolutionary exploits - they were using AI to accelerate routine coding tasks. The scam operations weren't creating novel fraud schemes - they were eliminating linguistic errors that previously exposed them. This mundane integration is precisely what makes it dangerous. It democratizes sophisticated capabilities, lowering barriers to entry while raising detection thresholds. #intelligence #AI #China #ChatGPT #OpenAI #disinformation

Foreign adversaries are no longer just targeting national policymakers—they’re infiltrating our local communities, elections, and digital spaces with sophisticated disinformation campaigns. A new report, by Krystyna Sikora and Bret Schafer "State(s) of Foreign Information Operations," lays out the latest tactics used by Russia, China, Iran, and others to sow discord, manipulate public opinion, and weaken trust in U.S. institutions. This report covers an operation in every state in the United States, and I would encourage you to take a look at the repeated and brazen interference in the US Information Environment. Here are five of the most eye-opening disinformation operations from the report: 🔹 Alaska Secession Hoax (2014) – Russian operatives created a petition on the official White House website calling for Alaska’s return to Russia. Bots amplified it, generating 39,000 signatures in days and later fueling fabricated media stories. 🔹 Iranian Hackers Pose as Proud Boys to Threaten Voters (2020) – Just before Election Day, Iranian cyber operatives sent emails impersonating the Proud Boys, intimidating Democratic voters into switching their support to Trump. The campaign was quickly exposed, but not before sowing fear and confusion. 🔹 AI-Generated Conspiracy About Maui Wildfires (2023) – China’s Spamouflage network used AI-generated images to spread the claim that the U.S. government deliberately started the fires as a military experiment. Fake accounts pushed the conspiracy in 31 languages, testing new influence tactics. 🔹 Deepfake Impersonation of a Ukrainian Official (2024) – A malign foreign actor (likely Russia) used AI-generated video and voice cloning to impersonate a top Ukrainian official in a Zoom call with Senator Ben Cardin—attempting to extract sensitive policy stances on Ukraine. 🔹 Louisiana Chemical Plant Explosion Hoax (2014) – Russian troll farms fabricated a breaking news event about a deadly explosion at a U.S. chemical plant, complete with fake witness testimonies, doctored videos, and bot-driven social media campaigns linking it to ISIS. 💡 The Takeaway: Disinformation is no longer a futuristic threat—it’s happening now. AI and social engineering are making influence operations cheaper, faster, and harder to detect. Whether in elections, protests, or crises, foreign actors are shaping narratives to divide and deceive. 🔍 So what can we do? ✔️ Improve media literacy – Recognizing tactics like AI-generated content and impersonation is key. ✔️ Strengthen cyber defenses – Social platforms, government agencies, and private sector orgs must collaborate to detect and counter these threats. ✔️ Stay informed – Reports like these provide critical insights into modern influence operations. 📢 How do you see AI changing the disinformation landscape? What steps should we take to protect our information ecosystem? Let’s discuss below. 👇 #Disinformation #Cybersecurity #AI #NationalSecurity

Important piece by former State Dept. Global Engagement Center Administrator James Rubin on how in shuttering the Center, the US unilaterally surrendered the global information space to authoritarian powers. Quote: "Our adversaries have spread lies to turn local populations against American policies across the globe, charging, for example, that U.S. drug companies were introducing diseases into Africa and that Americans were looting valuable Ukrainian artifacts. In Africa, Russia set up fake news organizations to channel false information to local media and bloggers, raising doubts about American vaccines. In Latin America, a news source created by Russia pumped out false stories purporting to come from Western newspapers. In Moldova, the Kremlin spread anti-Western propaganda to interfere in the country’s elections. Because these stories were being shared by local writers and broadcasters — some paid by Russia — they constituted a new and dangerous national security threat for the West." Quote: "Over time, together with international partners in Europe and Asia, we put in place the beginnings of a defense system. We were able to identify the hidden hand of our adversaries and inform governments, the private sector and the public that a particular series of news stories or claims were effectively dictated by the Chinese Communist Party or the Kremlin. Once aware of the source of news, listeners, readers and viewers could decide for themselves whether to believe the outrageous anti-American claims." https://wapo.st/4d5ehpj

What we talk about when we talk about #Disinformation? Perhaps we can learn from analysis of foreign influence operations, and direct those lessons inward to America. Two items that put information warfare from Russia and China into perspective must apply to our discussion of information tactics audiences see across the news in our own country. Wesley Winkler in Small Wars Journal (check the link in the comments below) presents some objectives of Kremlin- and CCP-backed information operations: ✅ Disrupts trust in institutions (Russia) ✅ Makes truth seem unknowable (Russia) ✅ Seeks to eliminate contradictions to create a monolithic reality (China) ✅ Operates through state-directed, top-down influence distributed through decentralized actor networks (China) If we follow the reporting on issues from the dismantling of #InternationalDevelopment assistance to the changes across international trade and tariff policy, is it possible that the current US administration is applying some of the same objectives to American political discourse? Winkler argues insightfully that #Russia and #China shape, distort, and dominate narrative landscapes to control perception, influence behavior, and assert strategic advantage. Sound familiar? Now bring into our discussion recommendations from this “Disinformation: A Measured Response” report by Jascha Heynen via the Center for Security Studies (CSS) at ETH Zurich: ➡️ Better methods of disinformation detection would allow decision-makers to devise more targeted solutions – and help them discern between orchestrated false narratives and legitimate criticism from within civil society. ➡️ Measures which increase trust in reliable news sources improve the quality of information environments more than those aimed at reducing the acceptance of false information. During the past three months, we have seen the current US administration eliminate foreign aid that improves the quality of #journalism by local outlets across the Global South, particularly in opposition to Russian and Chinese government-backed malign influence operations in countries where democracies struggle to survive. During these same three months, we have seen private sector #SocialMedia platform firms such as #Meta eliminate investments in fact-checking projects – what we know as simply professional journalism – that seek to identify false narratives and build trust in reliable news. Wesley Winkler hits the nail on the head: “Narrative control is most effective when people retreat into isolated, self-reinforcing tribes.” Yet again: Sound familiar? #MediaLiteracy #Misinformation #Propaganda #MediaDevelopment

One of the recurring mistakes when analysing Russian narratives is assuming they are primarily intended to convince external audiences. Very often, they serve multiple purposes simultaneously. They reinforce domestic narratives, justify future actions, create uncertainty, generate competing explanations and attempt to shape how events are perceived internationally. Facts become secondary to the narrative framework itself. This is why Russia frequently portrays itself as the injured party, even in situations where the broader context suggests otherwise. We have seen similar patterns in discussions around NATO enlargement, border incidents, alleged drone activity, sanctions, migration pressure and countless other issues over the years. The objective is rarely limited to explaining events. It is often about influencing how those events are interpreted. Understanding this distinction matters because responding only to individual claims risks missing the larger information environment in which those claims are being generated. Strategic communication, disinformation, influence operations and political narratives are not separate from the broader security environment. They are increasingly part of it. The challenge for democratic societies is therefore not only identifying falsehoods. It is maintaining the ability to distinguish between facts, narratives and deliberate attempts to shape perceptions at scale. https://lnkd.in/dZFx4dR7 #HybridThreats #Disinformation #Resilience #NationalSecurity

𝐓𝐡𝐞 𝐞𝐫𝐚 𝐨𝐟 𝐬𝐢𝐥𝐨𝐞𝐝 𝐩𝐫𝐨𝐩𝐚𝐠𝐚𝐧𝐝𝐚 𝐢𝐬 𝐨𝐟𝐟𝐢𝐜𝐢𝐚𝐥𝐥𝐲 𝐨𝐯𝐞𝐫. A fascinating new paper in Security Studies dissects how states have moved beyond simple "fake news" into what the authors term "𝐅𝐮𝐥𝐥-𝐒𝐩𝐞𝐜𝐭𝐫𝐮𝐦 𝐏𝐫𝐨𝐩𝐚𝐠𝐚𝐧𝐝𝐚 𝐂𝐚𝐦𝐩𝐚𝐢𝐠𝐧𝐬." It is no longer just about a bot farm here or a broadcast there; it is about the architecture of coordination. The article’s most compelling insight—the "golden thread" for analysts—is the taxonomy of 𝐢𝐧𝐭𝐞𝐠𝐫𝐚𝐭𝐞𝐝 versus 𝐩𝐚𝐫𝐚𝐥𝐥𝐞𝐥 campaigns. By mapping media channels across two axes (overt vs. covert, and broadcast vs. social), the authors reveal how actors like the CCP and Russia’s RIA FAN leverage cross-citation to amplify narratives. It is a sophisticated, if unsettling, synergy where state media validates the troll, and the troll amplifies the state. I must commend the Renee DiResta and Josh Goldstein for this timely and prescient work. Their rigorous analysis of leaked documents provides a necessary framework for understanding this specific modality of influence. It is a meticulously researched read that adds significant nuance to our understanding of malign influence. However, while this "full-spectrum" media analysis is vital, we must be careful not to mistake the map for the territory. In the broader context of 𝐈𝐫𝐫𝐞𝐠𝐮𝐥𝐚𝐫 𝐖𝐚𝐫𝐟𝐚𝐫𝐞, Strategic Influence is not confined to media alone; it is an omni-channel endeavor that aggressively spans economics, technology, and civil society. To focus solely on the digital media layer is to miss the economic coercion and technological infrastructure that often underwrite it. As we dissect these integrated media campaigns, I am curious where you see the greatest vulnerability in the defense. Is it the inability to detect the 𝐜𝐨𝐨𝐫𝐝𝐢𝐧𝐚𝐭𝐢𝐨𝐧 between overt and covert channels, or our tendency to overlook the economic levers pulling the strings behind the scenes? #IrregularWarfare #NationalSecurity #InformationWarfare #Propaganda #Geopolitics #StrategicInfluence #Russia #China #SocialMedia #HybridWarfare https://lnkd.in/gFXj7vBg

Another update on #AI and #Disinformation / #FIMI: The The European Centre of Excellence for Countering Hybrid Threats (Heidi H.) droped abother report on the use of AI by #Russia and China. Here is the gist: Core argument: Artificial intelligence is not replacing traditional disinformation, but amplifying it. China and Russia use AI as a force multiplier to increase scale, speed, targeting precision, and cost-efficiency of FIMI. Key differences: •  China: strong domestic AI ecosystem → moving toward data-driven, personalized, and strategic influence •  Russia: weaker AI capabilities → relies on adaptation, improvisation, and open-source tools Chinese use of AI  •  AI + big data + platforms → precision influence at scale •  Strong state-tech integration enables systematic, long-term operations •  Evolution toward “algorithmic cognitive warfare” •  Integration of AI into a long-term, state-coordinated strategy, combining: •  state agencies + private tech companies •  centralized data access + AI tools •  propaganda + algorithmic influence -> Move towards micro-targeted influence 1. Personalization and data-driven targeting •  AI used to analyze user behavior and sentiment in real time •  Enables tailored messaging for specific individuals or groups •  Example: #GoLaxy system 2. Content generation at scale •  social media posts, comments, articles •  memes & political narratives •  multilingual content 3. AI Translations •  AI use for multilingual campaigns  •  Localized Content for different audiences 4. Synthetic media  •  AI-generated “news anchors” (Spamouflage) •  Fake audio recordings targeting Taiwanese elections •  AI voiceovers 5. Astroturfing & fake engagement •  networks of fake accounts coordinated with AI •  AI to avoid bot detection •  Goal: simulate organic public support 6. Algorithmic manipulation •  Flooding platforms with AI-generated content (dominate search results & suppress critical narratives) •  Influencer networks used boost visibility of pro-China narratives Russian use of AI  •  reliance on existing tactics •  focus on volume, chaos, and narrative disruption •  AI as a pragmatic add-on Key characteristics 1. Content amplification •  scale production of disinformation •  automate existing workflows 2. Synthetic media for deception and shock value •  emotional manipulation •  undermining trust 3. Fake personas •  AI profile pictures for fake accounts •  fake journalists with AI-websites 4. Information laundering •  networks of fake news sites mimic legitimate outlets o  plagiarize real articles o  translate and adapt narratives 5. Automation •  large volumes of posts across platforms •  synchronization of thousands of messages 6. Poisoning AI systems •  Experiment with long-term ecosystem manipulation, • false narratives appear in LLMs, Wikipedia & platform fact-checking systems •  influencing training data of AI models and search engines

Microsoft’s second Threat Intelligence Election report for the USA published today: BLUF: Russian efforts are focused on undermining U.S. support for Ukraine while China seeks to exploit societal polarization and diminish faith in U.S. democratic systems. 🇷🇺 For example, the actor Microsoft tracks as Storm-1516, has successfully laundered anti-Ukraine narratives into U.S. audiences using a consistent pattern across multiple languages. Typically, this group follows a three-stage process: 1️⃣ An individual presents as a whistleblower or citizen journalist, seeding a narrative on a purpose-built video channel 2️⃣ The video is then covered by a seemingly unaffiliated global network of covertly managed websites 3️⃣ Russian expats, officials, and fellow travellers then amplify this coverage. 4️⃣ Ultimately, U.S. audiences repeat and repost the disinformation likely unaware of the original source. 🇨🇳 China is using a multi-tiered approach in its election-focused activity. It capitalizes on existing socio-political divides and aligns its attacks with partisan interests to encourage organic circulation. 💻 🇷🇺🇨🇳China’s increasing use of AI in election-related influence campaigns is where it diverges from Russia. While Russia’s use of AI continues to evolve in impact, People’s Republic of China (PRC) and Chinese Communist Party (CCP)-linked actors leverage generative AI technologies to effectively create and enhance images, memes, and videos. 🤡 Audiences do fall for generative AI content on occasion, though the scenarios that succeed have considerable nuance. The following factors contribute to generative AI risk to elections in 2024: ✔️AI-enhanced content is more influential than fully AI-generated content ✔️AI audio is more impactful than AI video ✔️Fake content purporting to come from a private setting such as a phone call is more effective than fake content from a public setting, such as a deepfake video of a world leader ✔️Disinformation messaging has more cut-through during times of crisis and breaking news ✔️Impersonations of lesser-known people work better than impersonations of very well-known people such as world leaders Report: https://lnkd.in/d-DjesN6