Russia, China, and Iran are all sharing the same playbook for corrupting Western elections. The U.S. is on track to be the next target this November.
A troveof Russian documents leaked in May outlines Moscow’s plans to use AI to warp foreign decision making and sway elections. The Kremlin’s “Projects 2026“ strategy reveals Russia’s plot to influence upcoming elections in Europe and Israel. Their tactics are similar to those used by China and Iran for years, and what emerges isn’t a Russian innovation — it’s a shared authoritarian playbook.
Projects 2026 calls for the creation of an opinion leader database to monitor the social media profiles of 10,000 influential individuals across Europe, including the 100 most prominent opposition figures in France alone. This is precisely the model Chinese state-aligned company GoLaxy piloted against the United States in 2025. The firm reportedly built constantly evolving psychological profiles of 117 members of Congress and roughly 2,000 additional American opinion leaders.
Researchers at Vanderbilt University warned that GoLaxy appeared prepared to target these individuals with messaging aligned to Beijing’s priorities. By monitoring the social media profiles of Western leaders, China and Russia are able to map the leader’s target’s values, beliefs, emotional tendencies, and vulnerabilities, making their influence operations more tailored and impactful.
In addition to using AI to build psychological profiles, the authoritarian playbook calls for countries like Russia and Iran to flood their targets with tailored generative AI content. Projects 2026’s “AI News” initiative calls for producing more than 500 short videos to target French audiences across six social media platforms. This tactic mirrors Iranian information operations during the 2025 and 2026 wars with Israel and the United States, when pro-Iranian networks pushed AI-generated and AI-assisted videos tailored to different audiences across X and TikTok, which racked up millions of views. These countries are running highly productive propaganda engines.
The most durable piece of Projects 2026 isn’t the content at all — it’s the plan to create fictitious think tanks and research institutes that can hand Russian narratives a “legitimate information source,” making them more likely to be cited by real news outlets and large language models. The leak shows this already underway: A site called the “World Center for Strategic Studies,” registered in March 2026, published unsigned analysis aimed at Western think-tank audiences, with versions planned in German, French, and Spanish. The fake think tank doesn’t need to fool a discerning reader. It needs to get quoted, indexed permanently, and cited by the next AI model that goes looking for sources on the topic.
Iran ran a cruder version of this same play: Iranian state-sponsored cyber espionage group has repeatedly impersonated real institutions — such as the Royal United Services Institute, the Aspen Institute, and the Washington Institute — to steal credentials and plant material under a trusted name. Projects 2026 proposes something more sustainable: Don’t just hijack existing credibility, but manufacture new institutions and let them accrue it over time.
Projects 2026 also proposes building a real Israel-based research institute staffed by Israeli citizens with genuine academic credentials. This approach is reminiscent of the 2014 “Iran Expert Initiative,” where the Iranian Foreign Ministry officials attempted to court and influence Western think tank researchers to promote their perspectives and give them legitimacy. Emails leaked in 2023 showed that Iranian diplomats had worked with Western experts across 10 think tanks to aggressively tout the merits of the 2015 nuclear deal between Tehran and major world powers, formally known as the Joint Comprehensive Plan of Action.
In the age of AI, influencing think tanks is narrative-laundering at its most dangerous. Once a fabricated or state-aligned source gets cited by a legitimate outlet or indexed by a search engine, it can become part of the training and retrieval corpus that both AI systems draw upon.
While the plays are reusable, the targets are new, and for Russia, the target is European elections.
The leaked documents outline a plan to swing Armenia’s parliamentary vote through a Russian-diaspora outlet, and a “Mitteleuropa” initiative aimed at reshaping alignments in Hungary and Slovakia. This is where the surveillance, flooding, and laundering modules converge. To counter it, policy solutions must be swift and targeted.
First, allied governments should require disclosure of foreign funding and foreign-state links for any media outlet, think tank, or “research institute” operating ahead of an election in an at-risk state — closing exactly the gap that lets a “World Center for Strategic Studies” pass as domestic and independent. Research by the Foundation for Defense of Democracies (FDD) into Qatari money in American higher education shows what a real regime requires: disclosures reaching individual researchers, not just institutions; low reporting thresholds with real penalties, registration in the style of the Foreign Agents Registration Act for state-directed “institutes” masquerading as independent; and provisions against the kind of obstruction Qatar used to block Texas A&M’s funding records in court. Russia is proposing to build the same laundering vehicle in Europe that Gulf money has already normalized in American academia. The fix already exists — it just needs to be strengthened.
Second, platforms should both demonetize and remove unlabeled AI-generated political content during designated pre-election windows in at-risk states. X’s own emergency policy during the Iran war, which suspended digital influencers from the ability to monetize their content for 90 days on undisclosed AI-generated war footage, shows platforms already have the infrastructure to act fast when the stakes are high enough. But demonetization alone left the content up and circulating; independent researchers found AI-generated war footage kept flooding feeds regardless of the policy. Elections warrant a harder line: If it’s undisclosed AI-generated content during an active election period, it shouldn’t just lose ad revenue. It should come down.
Third, the U.S. should prepare for future attempts at election influence by boosting the protection mechanisms it already has in place. Currently, 31 states have passed election deepfake laws, but they are under legal scrutiny because they touch on domestic speech. Federal action targeting foreign actors rest on firm ground. The 2002 McCain-Feingold Act prevents electioneering communications by foreign actors, but this only applies to broadcast, cable, and satellite communications. It must be updated to explicitly include paid internet and digital advertisements. It should also close the gap on AI-generated political content and require labeling so that foreign deepfakes can’t hide behind current ambiguities.
Projects 2026 is not evidence that Russia has cracked something new. It’s evidence that Moscow, Beijing, and Tehran are now running the same AI playbook against the same democracies, taking turns testing which module works best. The future outlined in Projects 2026 may not be here yet, but objects in the mirror are closer than they appear.
The Cipher Brief is committed to publishing a range of perspectives on national security issues submitted by deeply experienced national security professionals. Opinions expressed are those of the author and do not represent the views or opinions of The Cipher Brief.
Have a perspective to share based on your experience in the national security field? Send it to Editor@thecipherbrief.com for publication consideration.
Read more expert-driven national security insights, perspective and analysis in The Cipher Brief
Facts Only
* Russian documents outline plans to use AI to warp foreign decision-making and sway elections.
* The Kremlin’s “Projects 2026“ strategy involves influencing upcoming elections in Europe and Israel.
* Projects 2026 calls for creating an opinion leader database by monitoring social media profiles of 10,000 influential individuals across Europe, including the 100 most prominent opposition figures in France.
* A Chinese state-aligned company piloted a similar model against the United States in 2025, profiling 117 members of Congress and approximately 2,000 American opinion leaders.
* The strategy involves flooding targets with tailored generative AI content, such as "AI News," calling for over 500 short videos targeting French audiences across six social media platforms.
* The plan involves creating fictitious think tanks and research institutes to provide legitimate information sources for Russian narratives.
* An Iranian cyber espionage group has impersonated institutions like the Royal United Services Institute and the Aspen Institute to steal credentials.
* Projects 2026 proposes building a real Israel-based research institute staffed by Israeli citizens with academic credentials.
* The strategy involves creating legitimate sources for narratives, making them likely to be cited by news outlets and large language models.
* A plan exists to swing Armenia’s parliamentary vote through a Russian-diaspora outlet and a “Mitteleuropa” initiative affecting Hungary and Slovakia.
* Policy solutions suggested include requiring disclosure of foreign funding for media and think tanks, demonetizing unlabeled AI content during election windows, and updating U.S. laws on digital advertising to include AI content.
Executive Summary
Russia, China, and Iran share strategies for corrupting Western elections. Leaked Russian documents detail plans to use AI to influence foreign decision-making and elections, including a strategy called "Projects 2026." This strategy involves creating opinion leader databases by monitoring social media profiles to map values, beliefs, and vulnerabilities of influential individuals across Europe. The method involves flooding targets with tailored generative AI content, such as short videos, to influence public opinion. Furthermore, the plan seeks to manufacture legitimate information sources, like fictitious think tanks, to ensure narratives are cited by real news outlets and AI models. This process mirrors past tactics used by Iran and China in information operations.
The documented strategies involve leveraging AI for psychological profiling of leaders and deploying tailored content generation. The ultimate goal appears to be creating seemingly legitimate, state-aligned sources to influence public discourse, moving from hijacking existing credibility to manufacturing new institutional legitimacy. The proposed actions include manipulating political outcomes in specific regions, such as Armenia's vote, and reshaping regional alignments like the "Mitteleuropa." Countermeasures suggested involve requiring disclosure of foreign funding for media and research entities, demonetizing unlabeled AI content during election periods, and strengthening U.S. protections against foreign digital influence.
Full Take
The core dynamic observed is the convergence of advanced technological capability—specifically generative AI—with long-standing geopolitical objectives in information warfare. The shared playbook among Russia, China, and Iran suggests that the innovation lies not in the specific technological tools, but in the systemic application of psychological profiling, content flooding, and institutional laundering across a spectrum of targets. The move from hijacking existing credibility to manufacturing new, AI-indexed institutions represents an evolution from overt espionage to deep structural manipulation, using the emergent capabilities of large language models as the ultimate vector for influence, regardless of the specific state actor driving the operation.
The implication is that cognitive sovereignty is threatened not just by direct disinformation, but by the systemic corruption of the informational infrastructure itself—the very mechanisms through which reality is validated and sourced. The proposed countermeasures focus on transparency (disclosure), content control (demonetization/removal), and legal modernization (updating libel/advertisement laws). However, the resilience against this approach depends less on reactive policy and more on embedding a systemic skepticism that rejects the premise that any constructed source or data point can be assumed legitimate without verifiable provenance. The pattern suggests an escalation where actors leverage technical advancements to achieve older, deeply entrenched goals of narrative control, demanding defenses that address institutional opacity alongside digital content regulation.
Bridge questions: How can international frameworks effectively enforce transparency regarding academic and media funding when state-aligned entities operate under the guise of independent research? What is the long-term impact on public trust when AI-generated narratives become indistinguishable from legitimate sources in the information ecosystem? What new forms of accountability are necessary when influence operations target abstract concepts like "values" and "beliefs" rather than concrete facts?
Sentinel — Human
The text functions as a sophisticated synthesis of leaked material and geopolitical parallels, presented with strong argumentative structure, indicating human analysis built upon complex data rather than pure generation.