State of Information: Week of March 15–22, 2026

Three leaders claimed victory in the same war this week. All three were believed.

Trump said the US was "winding down." Iran's Mojtaba declared the enemy defeated. Netanyahu announced Iran's nuclear program was finished. Each claim circulated inside its own media ecosystem. None was challenged by the others.

That's the state of information in week three of the Iran war. Not a contest between truth and lies. A world where incompatible facts coexist, each supported by its own evidence base, each invisible to the audiences consuming the others.

This is the weekly report.

The Deepfake Flood Crossed a Threshold

The New York Times identified more than 110 unique pro-Iran deepfakes in the two weeks ending March 20. The Foundation for Defense of Democracies confirmed the number. The Foreign Affairs Forum called it a "qualitative threshold" — AI-generated content now makes up a larger share of war disinformation than traditional manipulation.

The fakes cover everything. AI-generated videos show Israeli civilians cowering as explosions rip through Tel Aviv. Fabricated footage depicts American warships being sunk by torpedoes. Iranian state-linked actors produced deepfakes of downed US fighter jets paraded through Tehran. The Handala group released new deepfakes of Netanyahu and former PM Naftali Bennett.

What changed this week isn't the existence of deepfakes. It's the ratio. Synthetic content has overtaken real manipulation as the primary disinformation vector in a major conflict. That's never happened before.

Platform response has been negligible. AFP confirmed AI fakes persist on X despite policy crackdowns. Meta's Oversight Board found the company's deepfake moderation "falls short" during fast-moving conflicts. The board's specific criticism: Meta's AI Info workflow is "too slow and too dependent on users disclosing synthetic media before moderators escalate a review."

Social media companies have done little since OpenAI released Sora last year, which made realistic video generation available through a simple app. The Iran war is the first major conflict fought in a post-Sora information environment. Platforms weren't ready. They still aren't.

Russia Proposed Staging an Assassination in Hungary

The Washington Post reported on March 21 that Russian operatives proposed "the Gamechanger" — a staged assassination attempt on PM Viktor Orban to rally voter support ahead of Hungary's April 12 election.

This sits alongside VSquare's earlier reporting: at least three GRU military intelligence officers are in Budapest, running troll farms, vote-buying networks, and propaganda campaigns. The operation is overseen by Sergei Kiriyenko, Putin's First Deputy Chief of Staff — the same official who coordinated Russia's election interference in Moldova.

The EU activated its rapid-response election disinformation system. It has 44 signatories, including TikTok and Meta. It's voluntary.

Here's the gap: the EU's tools are designed for social media manipulation. Russia has boots on the ground. GRU officers operating under diplomatic cover at the Russian embassy in Budapest aren't detectable by platform-level monitoring. TikTok flooding is one vector. Physical vote-buying networks run by military intelligence are another.

Opposition leader Peter Magyar publicly called Orban a "traitor" who "enlisted Russian agents to rig the vote." A government-affiliated security expert acknowledged the Russian interference was real and described it as without precedent. Polling shows a significant gap between government-affiliated and independent pollsters — itself potentially an information operation designed to obscure the race's actual status.

Twenty-one days to election. The EU has platform tools. Russia has intelligence officers. The mismatch is structural.

92 Million People in the Dark

Iran's internet blackout entered its fourth week. The New York Times confirmed 92 million citizens remain under near-total disconnection. Connectivity sits at roughly 1% of normal levels.

The government's own deadline — restoration by Iranian New Year (March 20) — passed without action. Instead, the SNSC issued an indefinite media gag order. Filterwatch, which published a leaked confidential government plan in January, reports this is permanent. State media and government spokespersons have confirmed unrestricted access won't return.

The IRGC Intelligence Organization is sending text messages to citizens warning that their activity on opposition Telegram channels is under surveillance. The Intelligence Ministry reported "multiple waves of arrests" targeting individuals accused of links to opposition networks or "information dissemination."

During Nowruz, security forces shot into crowds celebrating Chaharshanbe Suri (the fire-jumping tradition). The IRGC hanged 19-year-old wrestler Saleh Mohammadi in Qom on March 19 as a deterrence message. Footage leaked out only through VPN and Starlink — the last free-information channels inside the country.

The information warfare angle is asymmetric. Iran floods external platforms with deepfakes and sockpuppet campaigns while blocking outside information from reaching Iranians. The blackout doesn't just suppress dissent. It creates a one-way mirror: Iran can broadcast propaganda outward while preventing its population from accessing independent reporting about their own war.

IRGC-linked firms are replacing departing foreign telecoms, permanently nationalising internet infrastructure under military control. This isn't wartime censorship with a return date. It's a new architecture.

The Hormuz Factual Schism

The Strait of Hormuz became the week's sharpest information divide. Three media ecosystems reported three different realities about the same body of water.

US media: Iran is strangling global shipping. AP described an Iranian "stranglehold" on the strait.

Middle Eastern media: Iran is blocking only coalition ships. Al Jazeera reported the strait remains open to everyone else.

Asia-Pacific media: 89 ships are transiting per day, versus a normal average of 100–135. That number appears in no US reporting.

These aren't competing interpretations. They're incompatible factual claims. The Albis Perception Gap Index scored this story at 7.5 — the highest single-story PGI of the week. The US–Middle East regional pair gap hit 8.2.

A shipping firm reportedly paid Iran $2 million for Hormuz passage. Iran is developing a vetting system for transit. Whether this constitutes a "blockade" or "selective maritime control" depends entirely on which media ecosystem you trust.

The most important factual question — who is actually being blocked — is answered differently in each system. Readers don't disagree about what to think. They disagree about what's happening.

China's Domestic Information Campaign

Chinese state media ran a coordinated "US decline" narrative all week. The framing was consistent across Xinhua, Global Times, Guancha, and CGTN: the Hormuz crisis proves America has "the heart but not the strength."

Fortune reported Chinese media coverage frames Hormuz as proof that "US force meant to intimidate Beijing has instead punctured the illusion of US omnipotence." Asia Times documented how Chinese analysts view the war as a strategic windfall. Guancha ran a headline claiming "only Chinese ships and Iranian ships can transit Hormuz" — partially inaccurate based on AIS shipping data, but effective for domestic confidence.

A Taiwanese think tank, Doublethink Lab, revealed this week that a PRC state-affiliated firm compiled information on tens of thousands of prominent Taiwanese people, including 170 politicians, to support cognitive warfare and election interference campaigns. The Diplomat reported on a new Chinese "Terminator" information operation linking AI fear narratives to the Anthropic controversy, aiming to "increase political unrest" while undermining American AI development.

Meanwhile, PLA forces fired at least 27 rockets into waters near Taiwan over two days, explicitly describing the exercises as blockade preparation. The story got near-zero coverage outside Asia-Pacific. The world's information bandwidth was aimed at Hormuz.

The dual-track pattern: official diplomatic neutrality in public channels, coordinated "America is finished" messaging in domestic media. 1.4 billion Chinese citizens receive a systematically different picture of US military power.

Gulf States: The Mirror Operation

Abu Dhabi arrested more than 100 people for posting "misleading" videos of Iranian attacks. The Daily Mail, Livemint, and Forbes documented a sweeping crackdown across the UAE, Kuwait, and Qatar — tourists, expats, and influencers targeted under broad cybercrime laws for sharing content depicting war impacts.

The Committee to Protect Journalists documented arrests of journalists, interference with reporting, airstrikes damaging media infrastructure, and sweeping restrictions on coverage across the region.

This is the mirror image of Iran's campaign. Iran floods platforms with fabricated positive content. Gulf states suppress authentic negative content. Both produce the same outcome: a distorted information environment. One side manufactures fake reality. The other erases real reality.

For foreign nationals, the message is clear: sharing what you see can get you detained. The chilling effect extends well beyond those arrested.

The Pakistan-India Disinformation Campaign

The Hindustan Times exposed a coordinated Pakistan-linked disinformation campaign falsely claiming India's IRIS Dena warship was struck by the US. The campaign showed "moderate-to-high indicators of coordinated inauthentic behavior" including verbatim content duplication, synchronised posting patterns, rapid hashtag emergence, and abnormal engagement spikes.

The Iran war has become a surface for existing regional rivalries. The actual conflict serves as raw material for information operations that have nothing to do with Iran.

Armenia-France: A Dormant Account Wakes Up

CIVILNET reported a coordinated disinformation campaign targeting Armenia-France relations. A key account had been inactive since 2013 and resumed activity in early March 2026 — a signature of account hacking for influence operations. The campaign uses fabricated videos to discredit Armenia's strengthening ties with France since the 2020 Nagorno-Karabakh war.

The operation is small. But it illustrates how the global attention crisis — everyone watching Hormuz — creates cover for targeted campaigns in smaller theatres.

The Verification Crisis Deepens

This week confirmed what the Netanyahu deepfake incident first revealed: the verification infrastructure itself is broken.

Grok, Elon Musk's AI on X, called a real Netanyahu proof-of-life video "100% deepfake" while independent fact-checkers confirmed it was authentic. Three lawsuits are now filed against Grok for generating nonconsensual explicit AI content. The platform that hosts much of the Iran war's information is running an AI that both creates and misidentifies synthetic media.

India's Election Commission issued new rules requiring all AI-generated campaign content to carry labels covering 10% of visual or audio duration. The EU AI Act's Article 50 enforcement begins in August, mandating content labeling with penalties up to 6% of global revenue. India's 3-hour deepfake takedown rule — the world's most aggressive — is being tested in real time.

YouTube expanded its deepfake detection system to cover all US politicians, officials, and journalists. It uses a three-tier system: labeled synthetic content (allowed), unlabeled non-malicious deepfakes (requires labeling), and malicious deepfakes (immediate removal).

These are real responses. They're also reactive. The deepfake count is 110 and climbing. The platforms are playing defense against an offensive that costs almost nothing to mount.

What We're Watching Next Week

Hungary (21 days to election): The GRU staged assassination proposal adds a physical-security dimension to what was already the most aggressive Russian election interference operation in an EU member state. Whether Magyar's "traitor" framing cuts through Orban's fear-mongering about Ukraine will determine the narrative battle. Iran blackout (permanent?): The Nowruz deadline passed. IRGC infrastructure nationalisation is underway. The question is no longer whether the blackout lifts. It's whether Iran becomes a permanent digital island — and what that means for the 92 million people inside it. Deepfake escalation: The 110-count is two weeks old. At current production rates, the total likely exceeds 200 by now. Platform response remains inadequate. The Foreign Affairs Forum's "qualitative threshold" assessment means the information environment around this war has structurally changed. There's no going back to pre-Sora conditions. Natanz and the nuclear narrative: The US struck Iran's Natanz nuclear facility with bunker busters this week. Iran hit Dimona. Both sides called it self-defence. The nuclear dimension adds another layer of incompatible factual claims to an information environment already saturated with them. China-Taiwan: 27 rockets, blockade rehearsal language, near-zero global coverage. If the PLA escalates while the world watches Hormuz, the information gap will be measured in days, not hours.

---

The State of Information is a weekly Albis report tracking active information warfare campaigns, platform responses, and the structural forces shaping what the world sees — and doesn't see. Sources: New York Times, Washington Post, VSquare, FDD, Foreign Affairs Forum, Reuters, AFP, Balkan Insight, CPJ, CIVILNET, Hindustan Times, ISW, Doublethink Lab, The Diplomat, Fortune, Asia Times, Daily Mail, Meta Oversight Board, Filterwatch. Albis Perception Gap Index data from daily scan and PGI collection.