Trolling diplomacy: the evolution of digital statecraft from tweets to deepfakes

Engineering notes

Engineering notes will be added by the aipentium editorial team.

Chinese explanation / 中文解读

中文解读待补充：本站会优先为大语言模型、生成式AI、ChatGPT相关技术、计算机视觉、深度学习等高价值论文补充中文说明。

Original abstract

The digital transformation of diplomacy has entered a qualitatively new phase. A decade ago, the term "Twitter diplomacy" caught the attention of both scholars and practitioners, describing how political leaders bypassed traditional diplomatic channels to address a global audience directly through short messages (Bjola and Manor, 2024). That era, for all its audacity in shattering established norms, followed a recognizable logic: leaders sought to project carefully curated versions of reality, signal their intentions, and shape narratives through the sheer volume and speed of their communications.The Middle East conflict of 2025-2026 demonstrates that this model has given way to a fundamentally different one. The twelve-day war between Israel and Iran in June 2025 marked a turning point: for the first time, AI-generated disinformation constituted the largest category of false claims in a large-scale interstate conflict, surpassing even traditional propaganda in terms of volume and speed of dissemination (ISOC-IL, 2025). In the following year, both state and state-affiliated actors on both sides of the conflict employed deepfakes, AI-generated content, networks of fake accounts (bot farms), and deliberate ambiguity, not so much to persuade their audiences as to undermine the very possibility of a shared factual foundation. While "Twitter diplomacy" sought to disseminate, this new modus operandi aims to conceal. While the former was declarative, the latter is deceptive. This article introduces the concept of "trolling diplomacy" to provide a theoretical framework for understanding this shift. It is defined as a deliberate set of information operations, often involving the use of fake, generated, or contextually manipulated media content by state and state-affiliated actors, aimed at disrupting the adversary's information environment, provoking internal divisions, and creating strategic ambiguity.Unlike earlier forms of digital diplomacy, which operated within the liberal paradigm of "dialogue" and "engagement" (Manor, 2023), "trolling diplomacy" represents a manifestation of what Walker and Ludwig (2017) termed "sharp power" -the ability to infiltrate and manipulate the information ecosystems of rival states.In essence, this amounts to turning ontological destabilization into a weapon: its strategic advantage lies not in convincing the audience of a specific truth, but in making it impossible to establish the truth itself with certainty.Using the ongoing Middle East conflict as its primary case study, this article examines how two rival powers, the Islamic Republic of Iran and the United States, have put "trolling diplomacy" into practice. The analysis consists of four parts. First, the concept is situated within the context of existing literature on digital diplomacy, hybrid warfare, and propaganda. Second, the article describes the repertoire of "trolling diplomacy" techniques: AI-generated deepfakes, leaks, networks of fake accounts, and meme wars, as they manifested in the conflict. Third, the Iranian and American approaches are compared. Finally, it is argued that "trolling diplomacy" undermines the communication infrastructure on which the exchange of diplomatic signals and crisis management depend, thereby increasing the risk of unintended escalation.To understand "trolling diplomacy," we must first acknowledge that what we are witnessing is not so much an evolution of digital diplomacy as a fundamental transformation of it. "Twitter diplomacy" was a product of a specific technological and political moment -an era when social media was viewed as an instrument for democratization, facilitating "dialogue" between states and the international public (Bjola and Kornprobst, 2024). Whether it was used to broadcast carefully crafted narratives or to directly engage in provocative rhetoric, the underlying logic of communication remained the same: there was a message, there was a sender, and there was an audience that needed to be convinced of a particular version of reality. Zivo, 2025). All of this was done not in spite of the content's obvious artificiality, but largely because of it. The very ease with which these fabrications were exposed served Iran's purposes: it demonstrated to the domestic audience that the "resistance" was still capable of delivering symbolic blows, while simultaneously diverting attention from very real military setbacks on the ground. This transformation was made possible by the convergence of three factors. The first was the democratization of generative AI, which enabled state actors to produce convincing visual content on an industrial scale. As detailed in the MP-IDSA (2025) report on the Twelve-Day War, AI-generated content "turned the information sphere into a critical battlefield," allowing even relatively resource-constrained actors to flood the information space with synthetic media. Second, the architecture of social media platforms, where algorithmic amplification, targeting, and lax content moderation have created conditions under which synthetic content can reach millions of viewers before it is flagged as false. Third, strategic choice. Both Iran and the United States have transformed disinformation from a tactical support for military operations into an independent method of waging war. This is PSYOP`s in the traditional sense; it is what this article defines as "trolling diplomacy" -the systematic use of deceptive synthetic media content as a primary foreign policy instrument with its own strategic logic.Based on an analysis of the Twelve-Day War and its aftermath, four key techniques can be identified that make up the repertoire of "trolling diplomacy":1. AI-generated deepfakes as a strategic instrument. Unlike the sporadic use of fabrications in past conflicts, the current conflict demonstrates the institutionalization of deepfakes as a standard operational instrument. Iranian state media, including Press TV, have systematically disseminated AIgenerated video footage: From a slow-moving missile crashing into a Tel Aviv skyscraper (with the Google Veo logo clearly visible) to videos showing the Azrieli Towers ablaze (over 40 million views) and the destroyed Ben-Gurion Airport (FDD, 2025a).2. Proxy networks and plausible deniability. One of the central principles of "trolling diplomacy" is the blurring of the lines between official government messaging and grassroots activism. The Iranian operation "CAR ONLINE" serves as a prime example here: a Persian-language Telegram channel with over 400,000 subscribers distributed detailed instructions on creating fake accounts on X, using ChatGPT to translate messages into Hebrew, and mimicking the behavior of Israelis in order to spread demoralizing messages within Israeli society (FDD, 2025b). This method of "remote instruction" creates room for denial: the state does not publish disinformation directly; it merely establishes the infrastructure and methodology, which are then implemented by "independent" supporters. Added to this is the use of bot networks: Israel's Ministry of Diaspora Affairs documented a coordinated network of at least 100 accounts on X, which posted 241,712 messages during Operation Rising Lion, reaching millions of users (Israel Hayom/JNS, 2025).3. Strategic leaks and "controlled ambiguity". At the opposite end of the spectrum from the Iranian model, which relies on anonymity and forgeries, lies the American technique of "strategic leaks". In the run-up to the Israeli strike on June 13, 2025, the Washington Post (2025) reported that "widespread claims of disagreements between the Netanyahu government and the Trump administration regarding the strike on Iran were part of a coordinated disinformation strategy." Intelligence sources described the public rift as "entirely fabricated," intended to "lull Tehran into a false sense of security" on the eve of the attack. This is not the dissemination of false content, but the creation of a false context: the use of official and semi-official channels, traditionally associated with credibility, to systematically distort the opponent's perception of reality.4. Meme wars and the exploitation of cultural divides. Finally, "trolling diplomacy" actively draws on the arsenal of internet culture, adapting forms of digital folklore to achieve strategic goals. Following the launch of Operation "Epic Fury" in February 2026, the pro-Iranian group "Akhbar Enfejari" launched a campaign using AI-generated Lego-style cartoons depicting Trump and Defense Minister Hegset in satirical, compromising situations (Counterterrorism Group, 2026). This content, which aimed at an English-speaking audience and distributed across a variety of platforms, exploits existing cultural and political divisions within American society in an effort to undermine domestic support for military action.A comparative analysis of the approaches taken by the two main antagonists in the conflict reveals both fundamental differences in their operational logic and an alarming symmetry in the outcomes they produce.As the cases examined show, the Iranian model is characterized by a focus on asymmetric decentralization and public engagement. Iranian "trolling diplomacy" functions as a classic "weapon of the weak": unable to compete with the U.S. and Israel in the conventional military sphere, Tehran compensates for this with a massive presence in the information sphere. It is distinguished by three key features:1. A high degree of deniability. The use of proxy networks, fake accounts, and grassroots mobilization (as in Operation CAR ONLINE) allows Iranian leadership to distance itself from specific disinformation campaigns, even when the connection to state institutions is obvious to analysts. This model reflects the overall architecture of Iran's security strategy: a sprawling network of proxy actors (from Hezbollah to cyber groups) provides operational flexibility while maintaining strategic control (IRGC Cyber Doctrine, cited in ISPI, 2026).2. Focus on domestic and regional audiences. A significant portion of Iran's efforts is directed not at global public opinion, but at maintaining a "strategic narrative" within the country and among its allies in the "Axis of Resistance." AI-generated images of destroyed Israeli cities serve a compensatory function: they create a virtual reality of military successes that contrasts with the actual results on the battlefield. The American model, by contrast, relies on control over the information chain and the management of escalation. The U.S. approach is primarily "complementary" rather than "compensatory": information operations serve to support conventional actions rather than replace them. Key features:1. A focus on official and semi-official channels. Unlike the Iranian model, American "trolling diplomacy" rarely resorts to anonymous fabrications. Instead, it uses targeted "leaks" to the mainstream media, official briefings, and statements to create a distorted picture of its intentions in the eyes of the adversary. The case of the fabricated rift between Netanyahu and Trump is a textbook example of this.2. Institutional doctrinality. While Iranian operations are often opportunistic in nature, the American approach is embedded within a formal doctrinal framework. The updated U.S. Marine Corps publication on "Deception" (MCTP 3-32F, 2024) explicitly describes "the use of the information environment to deceive the enemy" as a standard operating procedure.Similarly, the Joint Doctrine for Information Operations (JP 3-13) provides for "the integrated use of information capabilities to influence, disrupt, or subvert the adversary's decision-making process".3. Vulnerability to the "boomerang effect". Paradoxically, it is precisely this reliance on official channels that makes the American model vulnerable to counter-narratives. When the Washington Post exposed the coordinated disinformation campaign, it not only undermined trust in that specific operation but also cast a shadow over future statements by officials, creating precisely the "strategic ambiguity" that "trolling diplomacy" seeks to generate.What both models have in common, however, is a shared effect: the systematic blurring of the line between fact and fiction, signal and noise, diplomacy and deception.The analysis conducted demonstrates that "trolling diplomacy" is not a marginal aberration, but rather an increasingly institutionalized feature of modern diplomacy.The Iranian and American cases reveal different operational logics but converge in a single destabilizing effect: the systematic degradation of the communication infrastructure upon which the exchange of diplomatic signals, crisis management, and the international legal order depend.This article argues that we are witnessing a fundamental shift in digital diplomacy.Where there was once "Twitter diplomacy" -direct first-person communication, "trolling diplomacy" is now taking shape: the systematic use of synthetic media content, disinformation, and calculated ambiguity as instruments of foreign policy.While the first model was driven by the logic of persuasion, the second operates on the logic of ontological destabilization. Its goal is not to make the opponent believe the correct version of events, but to deprive them of the ability to believe in anything at all. However, it would be a mistake to consider this situation unsolvable. Recognizing the phenomenon of "trolling diplomacy" as a distinct strategic practice is the first necessary step. Further steps should include the development of new norms governing the acceptable limits of synthetic media use in interstate communication.Just as the nuclear age gave rise to a complex architecture of arms control and confidence-building measures, so too must the era of "trolling diplomacy" give rise to its own mechanisms of deterrence and verification.

Links and sources

• PDF from original source

Comments