The Perception Gap Index dropped to 5.35 on March 29, scanning 63 stories across seven regions. The day's most fractured story wasn't a bombing or a battle — it was a chip bill. The US House Chip Security Act scored PGI 7.58, outstripping 3,300 war dead at 7.15 and the Houthi missile barrage at 7.13. The geopolitics tributary held at Competing Realities (6.08). The US ↔ Middle East pair remained the day's widest fracture at 6.54. The world cooled half a point from yesterday's 5.88 — not because narratives converged, but because the war's second month is grinding divergence into routine.
That shift from crisis to routine is itself the story.
A chip bill fractures harder than a war
The day's highest PGI didn't belong to any bomb, missile, or casualty count. It belonged to the [Chip Security Act](/blog/deepseek-chip-security-act-criminal-national-hero-2026) — a US House bill triggered by DeepSeek's breakthrough and a probe into whether Nvidia's CEO misled Congress about chip smuggling.
Reuters led with investigation: lawmakers questioning whether Jensen Huang's testimony "misled regulators" about the risk of AI chips reaching Chinese military universities. A $2.5 billion smuggling scandal. A national security emergency.
The South China Morning Post led with vindication: a US government panel "credits China's AI edge to open-source models and manufacturing dominance." Proof that export controls failed. The same panel's findings — literally the same document — are a damning investigation in Washington and a victory lap in Beijing.
The PGI hit 7.58 because the causal stories are irreconcilable. In the US version, Chinese AI progress happened despite America's containment strategy, through illicit channels. In the Chinese version, it happened because of China's own innovation, proving containment was always futile. One frames achievement as theft. The other frames investigation as jealousy.
The AP ↔ US pair scored 7.50 on this story alone. Two tech superpowers are looking at the same congressional report and reading mirror-image conclusions.
3,300 dead. The name gap.
Iran's war dead passed 3,300 on March 29, with nearly half confirmed as civilians. PGI: 7.15. The ME ↔ US pair: 8.00.
The number is the same everywhere. What changes is whether those dead have names.
Middle Eastern outlets centred individual stories: a family of four killed in Bushehr, a water treatment facility hit in Khuzestan. Al Jazeera's Arabic service named victims. It described neighbourhoods. It quoted survivors. The dead are martyrs (شهداء) — a word that carries theological weight, placing them in a framework of sacrifice and moral meaning.
Western outlets reported aggregate numbers. "More than 3,300 killed." The passive construction — killed by whom? — does work here. AP placed the death toll in a sentence about military operations, sandwiched between missile counts and Hormuz shipping disruptions. The civilians aren't characters in the story. They're a variable in a strategic equation.
The word "martyrs" doesn't translate into English. Not because there's no equivalent — but because using it would signal editorial alignment with one side. So the dead get no word at all. They're numbers. The gap between "martyrs" and "statistics" isn't just linguistic. It determines whether 3,300 dead generate outrage or indifference.
Three journalists in a marked car
Israel struck a vehicle clearly marked as a press car in Lebanon on March 29, killing three journalists. PGI: 6.98. The highest regional pair: ME ↔ US at 8.50 — the single widest gap of any story today.
Al Manar called one victim an "icon of resistance media." Arabic-language outlets published full biographical profiles — names, careers, family circumstances. The killing was framed as deliberate targeting: a marked press vehicle, well-documented coordinates, a strike pattern consistent with [previous attacks on journalists](/blog/ai-deepfake-coverage-iran-war-selective-reporting-2026) in the conflict.
US outlets reported it as an "incident under investigation." They barely named the victims.
The 8.50 gap measures something specific: not whether the strike happened, but whether these were people or a footnote. In one information ecosystem, three journalists died as individuals with stories worth telling. In another, an incident occurred. The investigation is pending.
This is D6 — Cui Bono divergence — at its sharpest. Naming the dead builds a case for war crimes prosecution. Not naming them preserves the frame of legitimate military operations with regrettable collateral damage. Each approach serves the legal and political interests of the region telling the story.
The Houthi question: terrorist or resistance
The Houthi missile barrage on Israel scored PGI 7.13 with the ME ↔ US pair at 8.00. The divergence is entirely structural.
AP: "Iranian-backed Houthi rebels entered the monthlong war, claiming two missile launches at Israel." BBC: 400 missiles fired, 92% intercepted, Dimona proximity raised. Al Jazeera: the strikes placed within a comprehensive "first four weeks" timeline of the "US-Israeli war on Iran."
The word "aggression" (عدوان) dominated Arabic-language coverage. It was absent from English coverage. Not because English has no word for it, but because applying it to the Houthis would require accepting their premise — that they're responding to something, rather than initiating something.
"Iranian-backed rebels" versus "resistance fighters responding to aggression." The same armed group performs opposite functions depending on your continent. This is the most stark hero-villain reversal in today's coverage, unchanged from [yesterday's parallel](/blog/pgi-breakdown-march-29-2026-same-data-five-alarm-levels) on the same actors.
Russia's quiet windfall
While the Iran war dominated attention, Russia pocketed €7.7 billion in oil windfall profits. PGI: 6.90. The framing split three ways.
Reuters framed Ukraine as resourcefully compensating for diplomatic losses — "using long-range strikes to maintain pressure after sanctions eased." The story is about Ukrainian agency.
The Guardian framed it as cascading devastation — "the largest supply disruption in the history of the global oil market." The story is about global harm.
The South China Morning Post asked the question no Western outlet would: "Is Russia really shaping up as the biggest winner from the Iran war?" The SCMP framed the windfall as evidence of American strategic folly — a war that "relatively favourable for Russia's geopolitical and geoeconomic standing." Beijing doesn't need to say the US is losing. It just needs to point at who's winning.
The interest-alignment: US framing keeps Ukraine sympathetic and worthy of support. EU framing builds the case for energy diversification. Chinese framing documents how American wars benefit Moscow — an implicit argument that US global strategy is self-defeating.
The River System: where the fractures run
The seven tributaries show a war that dominates the river while quieter crises flow beneath.
Competing Realities (red zone). Geopolitics (PGI-GP) at 6.08 — the only tributary above the Competing Realities threshold for the second straight day. Nineteen stories fed this stream. Iran's dead, the Houthi front, Kuwait's grounded flights, Oman's shuttered port, wounded Americans at a Saudi base, US Marine deployments, and Trump pushing his Hormuz deadline to April 6. The geopolitics stream isn't just running red. It's the river's main current.
Diverging Narratives (orange zone). Five tributaries sit here. Info Warfare (PGI-IW) at 5.62, pulled up by the journalist killing and AI-generated war propaganda going viral faster than real footage. Economics (PGI-EC) at 5.49 — Russia's windfall, Goldman cutting India's growth outlook, and fertilizer prices up 35% as Hormuz chokes global urea supply. Health (PGI-HE) at 5.25 from nine paramedics killed across five attacks in Lebanon. Technology (PGI-TE) at 4.82 — the DeepSeek story dragged the stream up despite the tributary's breadth. Climate (PGI-CL) at 4.78 with fifteen stories, the widest tributary by volume, anchored by India now getting 40% of its oil from Russia.
The calmest stream. Women's Rights (PGI-WR) at 4.12 with only two stories. Four hundred child marriages reported in Gaza — families selling daughters amid the siege — scored 5.03. A Bangsamoro fatwa against forced marriage scored 2.30. Low PGI on women's rights doesn't mean global agreement. It means most of the world isn't covering these stories at all.
The pattern holds from yesterday: proximity to the war's blast radius determines the perception gap. But something shifted. The geopolitics stream dropped a full point from yesterday's 7.02 to 6.08. Not because the war cooled — the dead are still climbing, the missiles still flying. Because the narratives are hardening. Each region has picked its frame and is repeating it. Divergence that was fresh on day one becomes routine by day thirty.
Cui bono: interest-alignment across regions
Narratives, like markets, follow the incentives of their producers. Today's stories illustrate the mechanism with unusual clarity.
The chip bill. US coverage of the Chip Security Act serves China hawks and export control advocates. The investigation angle pressures Nvidia specifically and the semiconductor industry generally — comply harder or face consequences. Chinese coverage of the same bill serves Beijing's tech narrative: export controls are failing, open-source innovation is winning, and America's containment strategy proves the weakness of the containeriser, not the contained.
The journalist strike. Arabic-language coverage naming victims and detailing the deliberate targeting of a marked press car builds the legal case for war crimes prosecution. American coverage treating it as an "incident under investigation" preserves the framework of legitimate military operations with occasional civilian harm. Neither outlet is lying. Both are selecting evidence that serves their institutional positioning.
The Houthi narrative. Framing Houthis as Iranian proxies justifies broader US military operations across the resistance axis. Framing them as a resistance movement responding to aggression justifies regional solidarity. The choice of frame predetermines the policy conclusion.
Russia's windfall. Chinese media documenting Moscow's gains serves Beijing's meta-narrative: that American military adventures benefit America's adversaries. The SCMP doesn't need to argue against the Iran war directly. It just shows who the war is making rich.
Every region's information marketplace produces narratives that serve its domestic institutions, economic structures, and geopolitical positioning. The interests don't create the divergence from nothing — they select which true facts to amplify and which to let fade.
Global Attention Index: the world sees prices but not people
The GAI hit 4.89 on March 29: Selective Visibility. Down 0.44 from yesterday. The global information map reveals a world that can read a commodity chart but not a casualty list.
The only universal story. Oil at $111 — up 55% in one month — achieved GAI 0.45. All seven regions covered it. The only story on the planet today that every major information ecosystem saw. It's raw market data: no hero, no villain, no moral framework. That's why it broke through. Numbers are the one language that translates without friction.
The attention desert. Information Warfare (GAI-IW) sits at 5.89 — the deepest attention gap among all tributaries. Stories about AI propaganda flooding social media, militaries building synthetic people for war content, and deepfake footage going viral faster than real footage are seen almost exclusively by US and EU audiences. The populations most targeted by these campaigns are the least likely to know they're being targeted.
The invisible crises. 3.2 million people displaced in one month of the Iran war: invisible to 5.38 billion people, including Americans whose military initiated the conflict. Sudan's confirmed famine: seen by EU and African outlets, invisible to the other five regions — 4.39 billion people. 52.8 million facing hunger across West Africa and the Sahel: visible almost exclusively in French-language media. The English-speaking world has functionally turned away from the largest food crisis on the planet.
Region blindness. Africa missed 90% of today's 63 stories. Latin America missed 88.3%. At the other end, the US missed 30% and the EU missed 28.3%. The information system isn't just unequal. The gap between the most and least informed regions is three-to-one.
The PGI × GAI crossover. The DeepSeek chip story scored PGI 7.58 but GAI 5.22 — high fracture, moderate visibility. Only the US and Asia-Pacific saw it with real depth. The rest of the world missed the most fractured story of the day. Compare that to the Houthi missile barrage: PGI 7.13 and GAI 3.03 — high fracture but broad awareness. Five regions covered the Houthis, creating abundant material for narrative disagreement. The chip story fractured harder but had fewer witnesses.
Then the starkest case: Iran's 3.2 million displaced scored GAI 6.73 — nearly invisible to the English-speaking world — but a PGI of just 6.23. Where the displacement story was covered, there was meaningful disagreement. But mostly, it wasn't covered at all. The perception gap isn't always between framings. Sometimes it's between seeing and not seeing.
The ME ↔ US pair: the defining fracture
For the second straight day, the Middle East ↔ United States pair sits atop the divergence table at 6.54. Twenty-six stories separate these two information ecosystems.
On every story touching the Iran war directly, the framing differences are architectural. Not a matter of emphasis. A matter of premises. In Washington, the war is a military operation with diplomatic off-ramps. In Doha, it's the "US-Israeli war on Iran" — a categorisation that assigns origin, blame, and moral responsibility before the first fact is reported. These two frames can't be reconciled because they disagree about what the story is.
The EU ↔ US pair, by contrast, sits at a manageable 3.40 across 41 shared stories. Europe and America read the same basic story — they argue about interpretation, not identity. "Weeks not months" in Washington versus "repeated NPT violations" in Brussels is a disagreement between partners. "Military operation" versus "war on Iran" is a disagreement between worlds.
What 5.35 tells us
The PGI dropping from 5.88 to 5.35 doesn't mean convergence. It means calcification.
Day thirty of the Iran war. The novelty has worn off. Each region has its frame, its vocabulary, its moral architecture. American outlets say "Iran war." Arabic outlets say "US-Israeli war on Iran." European outlets focus on civilian costs and nuclear safety. Asian outlets track the energy cascade. African outlets — when they see the war at all — track the food and fuel prices hitting their populations hardest.
These frames aren't shifting anymore. They're setting. A lower PGI at this stage of a conflict can mean the same divergence is simply less surprising. The perception gaps that scored 8.0 in week one score 7.0 in week four because the gap itself is priced in. Readers have already sorted themselves into information ecosystems. The walls between them are load-bearing now.
The day's sharpest lesson sits in a detail most readers will skim past: the Chip Security Act — a bill about semiconductor export controls — fractured global perception more than 3,300 war dead. PGI 7.58 versus 7.15. A chip debate outscored a body count because the war's deaths are now expected, categorised, filed under existing frameworks. The chip bill introduced a new question — who owns the future of AI — and new questions fracture harder than recurring tragedies.
Three journalists died in a marked press car. In one market, they were icons. In another, an incident. The gap between those two words — icon and incident — is 8.50 points of perception. It's also the distance between a war crime and a footnote.
The question worth sitting with: when does the unfamiliar fracture harder than the catastrophic? And what does it say about an information system where a congressional hearing splits the world more than a civilian death toll?