Information Warfare & AI Disinformation — Cascade Analysis¶

As of: March 24, 2026 (Day 24 of conflict) Built from: All country analyses, blind-spots/analysis.md, open-source reporting, disinformation tracking organizations

Core Thesis¶

The 2026 Iran War is the first major interstate conflict fought in a mature AI-disinformation environment. Every belligerent — the US, Israel, and Iran — is running parallel information campaigns, but the structural dynamics differ radically from any prior conflict:

AI-generated content now constitutes a larger share of conflict disinformation than traditionally manipulated content — a qualitative threshold crossed for the first time (EDMO, March 2026).
The US government is simultaneously waging information war abroad and restricting domestic press access to its own operations — creating an information vacuum that adversary propaganda fills.
Iran's internet blackout (99% connectivity loss since Feb 28) means the war is unfolding without the amateur civilian footage that shaped every conflict since the 2011 Arab Spring — the information terrain is dominated by state actors on all sides.
Algorithmic amplification on TikTok, X, and Instagram turns AI-generated fakes into mass-consumed narratives faster than any verification infrastructure can respond.

The cascade: information control shapes domestic opinion → domestic opinion constrains or enables military options → military outcomes generate new information → cycle repeats. Whoever controls the information tempo controls the war's political clock.

Capabilities Assessment: All Sides¶

United States¶

Offensive capabilities — HIGH: - US Cyber Command and NSA conducted the opening salvo of the war in cyberspace, disrupting Iranian communications and sensor networks before kinetic strikes landed (Flashpoint, March 2026). - Secretary Hegseth renamed Military Information Support Operations back to "Psychological Operations" (PSYOP) in December 2025, signaling doctrinal emphasis on influence warfare (DoD memo, December 8, 2025). - US/Israeli operations hacked Iranian prayer apps to push defection messages to millions of phones during the Feb 28 strikes. Major Iranian news outlets (IRNA, ISNA, Tabnak, Asr-e Iran, Rokna) were simultaneously taken offline (NPR, March 10, 2026). - Covert program smuggled thousands of Starlink terminals into Iran to maintain information access for dissidents (NPR, January 15, 2026).

Domestic information control — UNPRECEDENTED for modern US conflict: - FCC Chair Brendan Carr threatened broadcast license revocations over networks' Iran war coverage. Trump endorsed Carr "looking at the licenses" of "Highly Unpatriotic News Organizations" (Washington Post, March 14, 2026; CBS News, March 2026). - Pentagon banned photographers from press briefings. Defense Secretary Hegseth revoked credentials of major outlet reporters; replaced them with MAGA-aligned media ("the next generation of the Pentagon press corps") (CNN, March 4, 2026). - Journalists from virtually every major American news outlet turned in Pentagon press passes en masse in October 2025 after new restrictive policy (CNN, March 21, 2026). - Federal judge (Senior US District Judge Paul Friedman) ruled the Pentagon's press restrictions unlawful on March 20, 2026 — but enforcement during active hostilities remains uncertain (CNN, March 20, 2026). - Net effect: The war is operating as a "black box" — Pentagon beat reporters cannot get answers to basic operational questions. No background briefings, no embedded journalists, no independent verification of military claims.

Vulnerability — CRITICAL: - The domestic information vacuum creates space for adversary narratives to fill. Without credible independent reporting, every claim becomes contestable. - FCC threats, while legally limited to local broadcast stations (not cable/streaming/print), create a chilling effect across the media ecosystem. - Historical parallel: The 2003 Iraq War showed that media docility in wartime produces catastrophic policy failures. The WMD narrative succeeded in part because pro-war sources outnumbered anti-war sources 6:1 on network news (FAIR, 2003). The current environment is more restrictive than 2003.

Israel¶

Offensive capabilities — VERY HIGH: - Unit 8200 (SIGINT/cyber) and Unit 9900 (visual intelligence) provided the targeting intelligence for the strike on Khamenei's bunker, including hacking Tehran traffic cameras and monitoring communication networks for over two decades (Jerusalem Post, March 2026). - IDF Spokesperson's Unit doubled to 200+ personnel after October 7, 2023, with multilingual capacity in 14 languages. - "Predatory Sparrow" (Gonjeshke Darande) group — linked to Israel — conducted cyber attacks on Iranian banking (Bank Sepah) and infrastructure (Euronews, March 2, 2026). - IDF established strong social media presence: Avichay Adraee built 2.5M Facebook followers, 554K X followers, and a comparable TikTok following for Arabic-language influence operations. - IDF operates Substack ("Mission Brief") and other direct-to-audience channels that bypass traditional media gatekeepers.

Vulnerability — MODERATE: - Israel's narrative depends on maintaining the "surgical precision" frame. Any verified mass-casualty event among Iranian civilians undermines this. - With Iran's internet largely down, Israel has limited ability to reach the Iranian domestic audience directly — the blackout cuts both ways.

Iran¶

Offensive capabilities — MODERATE but IMPROVING RAPIDLY: - IRGC's Seraj Cyberspace Organization (founded 2016) trained thousands of recruits in social media content production, hashtag campaigns, and coordinated inauthentic behavior. - Cotton Sandstorm (IRGC-affiliated) specializes in hack-and-leak operations, website defacements, and DDoS attacks combined with influence amplification. - Press TV (English), Al-Alam (Arabic), and 8 international IRIB channels provide state-controlled broadcast infrastructure. - Wikipedia cited IRGC-linked media over 78,000 times, indicating deep penetration of reference ecosystems (Jerusalem Post investigation).

AI disinformation campaign — THE DEFINING FEATURE: - New York Times identified 110+ unique AI-generated deepfakes conveying pro-Iran messages in the first two weeks of conflict alone (NYT, March 2026). - Cyabra (research firm) documented a pro-Iran campaign generating 145 million views and 9 million interactions across platforms within days (Cyabra, March 2026). - Tens of thousands of fake accounts deployed synchronously to disseminate AI-generated content (FDD, March 19, 2026). - Content types: deepfake videos of downed US aircraft, fake Western influencer personas supporting Iran, fabricated "betrayal" narratives about Iranian military figures (falsely claiming Quds Force commander Qaani arrested/executed), AI-generated battlefield imagery. - A single AI-generated video of an Iranian missile destroying a US fighter jet received 70 million views on X before BBC Verify traced it to a video game (Albis, March 2026; Rolling Stone, March 2026).

Critical vulnerability — SEVERE: - Iran's own internet blackout (99% connectivity loss) means the regime cannot reach its domestic population through digital channels. GPS jamming has cut satellite internet by 80%. Starlink users face up to 10 years imprisonment or execution under 2026 legislation. - The blackout also means no civilian footage from inside Iran — the world sees only what state actors on all sides choose to show.

Russia¶

Role: AMPLIFIER and EXPLOITER: - Coordinated Russia-China-Iran disinformation network exposed on X, operating through "Global Insight Journal" account created September 2025 and verified February 2026 — weeks before the war began (investigative reporting, March 2026). - Russian state media and information ecosystem amplifies Iranian AI-generated content, extending its reach into European and Global South audiences. - Russia uses the Iran war to redirect attention from Ukraine: "Is Russia using the war in the Middle East to spread disinformation about Ukraine?" (Euronews, March 17, 2026). - Kanopus-V/"Khayyam" satellite intelligence sharing with Iran provides targeting data — this has an information warfare dimension, as it allows Iran to make more credible operational claims.

China¶

Role: STRATEGIC NARRATOR: - Chinese state-aligned media echoes anti-US narratives, framing the war as evidence of "the US as a global troublemaker" — a core theme of China's strategic communications. - Targets Global South audiences with food and energy security narratives, positioning China as stabilizing force vs. US disruption. - Does not produce fabricated content at Iran's scale — instead curates, amplifies, and contextualizes within a coherent geopolitical frame. - Information operations reinforce China's real-world positioning as mediator and Gulf security alternative.

The AI Disinformation Environment: What Changed¶

Technical Capabilities in 2026¶

Capability	2003 (Iraq War)	2014 (Ukraine/ISIS)	2023 (Gaza)	2026 (Iran War)
Fake video	Crude edits	Basic compositing	Early deepfakes	Indistinguishable from real at social-media resolution
Voice cloning	Not possible	Not possible	Detectable	Crossed "indistinguishable threshold" — seconds of audio suffice (Fortune, Dec 2025)
Fake imagery	Photoshop	Photoshop at scale	Early AI generation	Convincing fake satellite imagery now documented (FlowingData, March 2026)
Scale	Manual	Bot farms (thousands)	Coordinated networks	Tens of thousands of accounts deploying AI-generated content simultaneously
Detection	Human review	Human review + metadata	Emerging AI detection	Detection infrastructure overwhelmed; platforms retreating from moderation
Deepfake volume	N/A	N/A	~500K globally	~8M globally (DeepStrike estimate, 2025 baseline; higher in wartime)

Key threshold crossed: "Deepfake-as-a-Service" platforms proliferated in 2025, democratizing the production of convincing synthetic media (Cyble, 2025). State actors no longer need specialized capabilities — the tools are commercially available.

Platform Dynamics¶

X (formerly Twitter): - Owner's "hands-off approach to moderation" means lowest removal rates for disinformation. - Only sanction for AI fakes: creators in payment program suspended for 90 days. The vast majority of users are unaffected. - Pro-Iran deepfake video reached 70M views before verification.

TikTok: - Algorithmic amplification reinforces emotional content through engagement-based personalization. Emotional war content is algorithmically favored. - Research shows TikTok's recommender system "repeatedly confirms and amplifies specific affective responses" — it does not merely show emotional content but creates feedback loops (EPJ Data Science, 2026). - TikTok's sale to Larry Ellison-led consortium sparked exodus to Palestinian-founded app UpScrolled, fragmenting the platform ecosystem.

Meta (Facebook/Instagram): - Did not respond to CNN requests for comment on Iran war AI fakes (CNN, March 11, 2026). - Post-2024 retreat from aggressive content moderation leaves gaps.

Net effect: Social media platforms have largely abandoned the aggressive moderation posture they adopted during the 2022 Ukraine conflict. The information environment is more permissive for state-sponsored disinformation than at any point since 2016.

The "AI-Native War" Framework¶

Researcher Erkan Saka (Istanbul Bilgi University) characterizes this as the first "AI-native war" — a conflict where AI-generated disinformation is not an add-on but a core feature of the information battlespace from Day 1. Key findings from his tracking (as of March 23, 2026):

Iran produces the highest volume of directly fabricated content (deepfakes, fake battlefield footage, synthetic personas).
US/Israel achieve propaganda effects through different means: narrative framing, strategic omission, media censorship, and delegitimization of independent journalism.
Critical insight: "Asymmetry in fabrication volume does not imply asymmetry in propaganda effect" — the cumulative epistemic damage may be equivalent even if the technical forms differ.

Iran's Internet Blackout: The Information Void¶

Iran's connectivity has been at approximately 1% of normal levels since the war began (The Register, March 17, 2026), building on a blackout that started January 8, 2026 during the pre-war protest crackdown.

What the blackout enables (for Iran's regime):¶

No amateur civilian footage of strikes, casualties, or infrastructure damage — the regime controls all imagery leaving the country.
No independent verification of Iranian military claims or casualty figures.
No domestic digital organizing against the regime during wartime.

What the blackout prevents (for Iran's regime):¶

Cannot reach domestic audience with its own digital propaganda — must rely on broadcast media.
Cannot coordinate distributed disinformation campaigns domestically.
Creates a credibility problem: if Iran claims victories but its population cannot verify, the narrative is fragile.

What the blackout enables (for US/Israel):¶

Iranian civilian suffering is invisible to global audiences — reduces political pressure for restraint.
US/Israeli narrative faces less real-time contradiction from ground-truth footage.
Smuggled Starlink terminals (~thousands covertly introduced by US) provide selective information channels to dissidents.

What the blackout enables (for the world):¶

An information vacuum is not an absence of information — it is an invitation for fabrication. Both sides fill the void with claims that cannot be independently verified.
Bloomberg: "Iran War: Internet Shutdown Limits Civilian Social Media Images" — this means the war's human cost is systematically invisible (Bloomberg, March 20, 2026).

Cascade Model: How Information Warfare Affects the War¶

Cascade 1: Domestic Opinion → Military Options¶

Pentagon press blackout
  → No independent war reporting
    → Public relies on government claims + social media
      → AI fakes fill information vacuum
        → Public opinion becomes unstable and manipulable
          → Political constraints on war policy become unpredictable

Current polling data (as of mid-March 2026):

Metric	Finding	Source
Approve strikes	40-43%	Emerson / Politico-PublicMe
Oppose strikes	47-53%	Emerson / Quinnipiac
Oppose ground troops	78%	Quinnipiac
Trump Iran handling approve	36-38%	Marist / Quinnipiac
Independent approve	24% (falling from 30%)	YouGov/Economist
Republican approve	79%	Multiple
Democrat disapprove	86%	Multiple

Key dynamic: Independent voters — the swing constituency — are moving against the war at an accelerating rate (-16 net swing in one week per YouGov). This is the number that drives Trump's political calculus. If the information environment shifts (e.g., a viral verified mass-casualty event), Independent opposition could collapse remaining support for continued operations.

Historical parallel: George W. Bush's Iraq War approval started at 72% (March 2003) and declined to 30% by 2008 as the WMD narrative collapsed. The current conflict started with much lower baseline support (40-43%), meaning the political runway for sustained operations is significantly shorter.

Cascade 2: AI Fakes → Verification Collapse → Policy Paralysis¶

AI deepfake of mass civilian casualties goes viral
  → 70M+ views before any fact-check
    → Congressional pressure for investigation
      → Pentagon cannot show independent evidence (press excluded)
        → "Is it real?" becomes unanswerable
          → Policy debate freezes around epistemological uncertainty

This cascade has already partially activated. The 70M-view fake video of an Iranian missile destroying a US fighter jet demonstrates the mechanism. A fake mass-casualty event — particularly one depicting US weapons killing Iranian civilians — would have significantly higher political impact than a fake military victory.

Scenarios of maximum disruption: - AI-generated deepfake of a US airstrike hitting a hospital, with fabricated satellite imagery as "evidence" - Voice-cloned audio of a US military official discussing civilian casualties (voice cloning has crossed the "indistinguishable threshold" per Fortune, December 2025) - Synthetic "leaked intelligence document" purporting to show US awareness of civilian harm - Any of the above amplified through the Russia-China-Iran coordinated network before verification is possible

Cascade 3: Iranian Diaspora → European Policy¶

250K protesters in Munich (Feb 14, 2026)
  → Largest Iranian diaspora protest in European history
    → 350K in Toronto, 350K in Los Angeles, 168 protests across 73 cities
      → European Parliament bans Iranian diplomats (Jan 12, 2026)
        → IRGC designated as terrorist organization by EU Council
          → European policy shifts from cautious neutrality toward active opposition to Tehran

This cascade is already well advanced, but the information warfare dimension adds complexity:

The diaspora protests are pro-regime-change, aligned with exiled Crown Prince Reza Pahlavi. They support the US/Israeli military campaign.
But European publics are anti-war (consistent with European polling since 2003).
European governments are caught between diaspora pressure (supporting regime change) and public opinion (opposing military action).
Iranian AI disinformation targeting European audiences aims to split this dynamic — portraying the war as disproportionate Western aggression while ignoring regime brutality.
Russian amplification of anti-war narratives in Europe serves the dual purpose of weakening European support for sanctions on Iran AND redirecting attention from Ukraine.

Cascade 4: Information Asymmetry → Global South Narrative¶

Iran's blackout → no civilian footage from inside Iran
  → US/Israeli narrative dominates Western media
    → But Russia/China curate alternative narrative for Global South
      → "US as global troublemaker" frame resonates in Africa, Asia, Latin America
        → UN General Assembly votes shift
          → Diplomatic isolation of US position
            → Reduced coalition support for sustained operations

This mirrors the Iraq War pattern but is accelerated by AI-generated content. Russia and China have particular success shaping narratives around food and energy security in Africa and the Middle East — tying the war to fertilizer prices, oil disruption, and food insecurity in ways that resonate with populations already experiencing these effects.

Cascade 5: Platform Fragmentation → Narrative Balkanization¶

TikTok ownership change (Ellison/Oracle)
  → User exodus to UpScrolled and alternative platforms
    → Information ecosystems fragment along political/ethnic lines
      → No shared factual baseline across audiences
        → Each population sees a different war
          → Coalition management becomes impossible

This is a structural shift without historical precedent. In 2003, there were three US broadcast networks and a handful of cable channels. In 2026, there are dozens of platforms, each with different moderation policies, algorithmic biases, and audience demographics. The result is not a "fog of war" but multiple simultaneous wars — each audience is experiencing a fundamentally different conflict.

Political Impact Modeling¶

Scenario A: Information Status Quo Holds (50% probability)¶

Pentagon maintains press restrictions despite court ruling.
AI fakes continue at current rate (110+ per two weeks) but no single event breaks through to mainstream consciousness with devastating effect.
Public opinion continues slow erosion among Independents.
Trump faces 30-35% Iran handling approval by May, constraining but not preventing continued strikes.
European policy remains split between diaspora pressure and public anti-war sentiment.

Scenario B: Viral Verification Failure (25% probability)¶

A convincing AI-generated mass-casualty deepfake goes viral and resists rapid debunking — either because verification infrastructure is overwhelmed or because the Pentagon's own press restrictions make independent confirmation impossible.
Congressional hearings demanded. "Did this happen?" becomes the defining question.
Independent voter support for war collapses to single digits.
Ceasefire pressure becomes irresistible within 1-2 weeks of the event.
This is the scenario where information warfare directly determines military outcomes.

Scenario C: Iranian Information Infrastructure Collapse (15% probability)¶

US/Israeli cyber operations or physical strikes degrade Iran's remaining broadcast capability.
Combined with internet blackout, Iran loses ability to project any narrative — even to its own population.
Regime faces internal legitimacy crisis: cannot communicate with citizens during wartime.
Could accelerate regime fragmentation OR trigger desperate escalation as the regime loses information control domestically.

Scenario D: Deepfake Escalation Spiral (10% probability)¶

AI-generated content triggers real-world military responses. Example: a fake intelligence report of an imminent Iranian attack causes a preemptive strike, which then triggers actual retaliation.
Or: a deepfake of a US/Israeli leader making threatening statements (voice-cloned) provokes Iranian miscalculation.
This is the nightmare scenario — the information domain causing physical escalation rather than merely reflecting it.
Probability is low but consequence is extreme. Detection infrastructure is not equipped to prevent this in real time.

Structural Observations¶

What Makes This Different From Every Previous Conflict¶

First war where AI-generated disinformation outpaces human-generated disinformation. The volume, speed, and quality of synthetic content exceeds what human operators could produce.
First war where the attacking power (US) is simultaneously restricting its own media's access to operations. In Vietnam, press freedom existed but undermined support. In the Gulf War and Iraq War, embeds were controlled but present. In 2026, independent media is effectively excluded from the Pentagon.
First major conflict with a near-total internet blackout in the target country. Iran's 99% connectivity loss means the information terrain is entirely controlled by state actors. There is no "citizen journalism" from inside Iran.
First conflict where platform moderation has been deliberately weakened relative to the previous major conflict (2022 Ukraine). X's hands-off policy, Meta's retreat from aggressive moderation, and the fragmentation of the TikTok audience mean there are fewer guardrails than in any conflict since 2016.
First conflict where voice cloning has crossed the indistinguishable threshold. A few seconds of audio can generate a convincing clone of any leader's voice, complete with natural intonation and breathing (Fortune, December 2025). This capability did not exist during the Gaza war of 2023.

The Iraq War Comparison¶

Dimension	Iraq 2003	Iran 2026
Pre-war narrative	WMD claims; fabricated evidence (aluminum tubes, Curveball)	Nuclear program claims; less contested factual basis
Media access	Embedded journalists (controlled but present)	Press effectively excluded from Pentagon
Source balance	Pro-war sources 6:1 over anti-war (FAIR)	MAGA-aligned outlets replace mainstream press corps
Disinformation technology	State-crafted talking points amplified by compliant media	AI-generated deepfakes at industrial scale by all parties
Public starting support	72% (March 2003)	40-43% (March 2026)
Verification capacity	Slow but functional (Judith Miller/NYT took years to debunk)	Overwhelmed in real time; 70M views before fact-check
Foreign amplification	Minimal	Russia-China-Iran coordinated network documented
Information fragmentation	3 networks + CNN/Fox	Dozens of platforms, algorithmic silos, no shared baseline

Key lesson from Iraq: The WMD narrative succeeded not because evidence was strong but because the media environment was structurally incapable of challenging it in real time. The 2026 environment is worse on every dimension except one — baseline public skepticism is higher, which provides some resistance to state narratives but also makes the public susceptible to adversary counter-narratives.

Key Uncertainties¶

Can detection keep up with generation? Current AI detection tools have high false-positive rates and struggle with compressed social-media-quality content. If detection remains behind generation, the information environment degrades indefinitely.
Will the court ruling on Pentagon press access be enforced during active hostilities? Judge Friedman's March 20 ruling voided Hegseth's press restrictions, but the administration may ignore or slow-walk compliance. Enforcement during wartime has no modern precedent.
What is the threshold for a "verification failure" that changes policy? A single viral deepfake has not yet forced a policy response. But the 70M-view fake fighter-jet video demonstrates the mechanism. A sufficiently convincing mass-casualty deepfake could cross this threshold.
How does the information war interact with the November 2026 convergence? (See cascades/november-2026-convergence.md.) Midterm elections will intensify every information warfare dynamic: candidates will use war footage (real and fake) in campaigns, FCC threats become more politically charged, and the pressure to control the narrative will peak. The information warfare cascade and the November convergence are deeply interlinked.
What happens when Iran's internet partially restores? If the war ends and connectivity returns, millions of Iranians will encounter weeks of accumulated disinformation from all sides simultaneously. The post-blackout information shock could destabilize the post-war political settlement.

Connections to Other Cascade Models¶

Combinatorial Matrix (cascades/combinatorial-matrix.md): Information warfare is not resource-dependent but amplifies every resource cascade — public perception of energy prices, food shortages, and chip supply is shaped by the information environment, not just underlying data.
November 2026 Convergence (cascades/november-2026-convergence.md): Midterm elections transform the information war from a military support function into a domestic political weapon.
Cyber Escalation (cascades/cyber-escalation-scenarios.md): Cyber operations and information operations are deeply intertwined — the same Unit 8200/NSA capabilities that conduct SIGINT also enable PSYOP. Hack-and-leak operations are simultaneously cyber attacks and information warfare.
Insurance Systemic Risk (cascades/insurance-systemic-risk.md): If a deepfake of a major maritime attack triggers insurance market panic, the economic cascade occurs regardless of whether the attack was real.

Sources¶