When Open-Source Synthesis Outperforms Classified Intelligence at the Tactical Level
The Fallacy
The classification system rests on a premise so deeply embedded in American defense culture that questioning it feels like questioning gravity: classified information is more valuable than unclassified information, and the architecture that protects secrets simultaneously protects the people who hold them. This is The Classification Fallacy. It confuses the protection of sources and methods—a legitimate and necessary function—with the protection of the force. These are not the same thing. They have never been the same thing. And on the seventh day of Operation Epic Fury, with six American soldiers dead in Kuwait and Iranian command-and-control fragmenting into uncoordinated retaliation, the distance between those two functions is measured in body bags.
The fallacy operates through a simple inversion. The system classifies information to keep it away from adversaries. But the architecture required to enforce that classification—compartmentation, need-to-know restrictions, echelon-based dissemination, and the sheer friction of moving cleared material through secure channels—simultaneously keeps information away from the very people the system was built to protect.
A specialist at Camp Arifjan knows what her battalion S-2 briefed twelve hours ago, filtered through classification restrictions, command messaging priorities, and whatever her commander decided was relevant to her lane. She does not know that Iran’s own Foreign Ministry admitted on March 3 that its military has lost control of several units operating on prior general instructions. She does not know that Iranian ballistic missile attacks have dropped ninety percent while drone hit rates have quadrupled—a shift that fundamentally changes her threat model. She does not know that the Strait of Hormuz is functionally closed, that CSIS estimates the first hundred hours of this operation cost $3.7 billion, or that the President of the United States demanded unconditional surrender from a decapitated regime whose surviving commanders cannot coordinate their own forces. All of this is open-source. None of it is classified. And she almost certainly does not have it.
This is not a new failure. It is the oldest failure in American intelligence, wearing new clothes. The Department of Defense Committee on Classified Information warned in 1956 that overclassification had reached “serious proportions.” A joint CIA-Department of Defense commission found in 1994 that the classification system had “grown out of control.” The 9/11 Commission concluded in 2004 that compartmentation contributed directly to the failure to detect the September 11 plot. The Reducing Over-Classification Act became law in 2010. And here we are in 2026, with the same architecture, the same culture, and six dead Americans in Kuwait who might have been better served by a twenty-three-year-old with a laptop and an Al Jazeera feed than by the most expensive intelligence apparatus in human history.
The Center of Gravity
The center of gravity is not the classification of any individual document. It is the synthesis architecture—or rather, the absence of one. The intelligence community generates enormous volumes of both classified and open-source material, but no echelon below combatant command is chartered, staffed, or equipped to fuse open-source streams across domains into real-time tactical intelligence products. The problem is not that the pieces do not exist. It is that the institutions holding the pieces are architecturally prevented from assembling them.
Government officials have conceded for decades that between fifty and ninety percent of classified documents could safely be released, a finding documented by the Brennan Center for Justice and confirmed by officials ranging from former Defense Secretary Donald Rumsfeld to former CIA Director Porter Goss, who told Congress that the intelligence community “overclassifies very badly.” The Reducing Over-Classification Act of 2010 codified what Congress had known since at least 2004: that the 9/11 Commission found “security requirements nurture over-classification and excessive compartmentation of information among agencies.” Sixteen years after that law, with fifty million classification decisions made annually, the architecture remains fundamentally unchanged. The ODNI’s own 2024 strategy document acknowledged that the office is “driving classification reform,” a phrase that would be encouraging if it had not been the same phrase used by every DNI since the position was created.
Meanwhile, former CIA officer Arthur Hulnick estimated that as much as eighty percent of the intelligence database is derived from open-source material, a figure cited by the Australian Army’s analysis of tactical OSINT application. The Defense Intelligence Agency published its 2024–2028 OSINT Strategy, and the ODNI’s own 2024–2026 OSINT Strategy stated that “the ability to extract actionable insights from vast amounts of open source data will only increase in importance.” The intelligence community knows the value of open-source material. It simply cannot deliver it to the echelon that needs it most.
The scale of the failure is staggering when measured against the resources deployed. Approximately 4.2 million Americans hold security clearances—nearly one in every fifty adults. The government spends billions annually on personnel security, classification management, and the physical infrastructure of secrecy: SCIFs, secure communications, cleared courier networks, and the bureaucratic apparatus required to process, store, protect, and eventually declassify the material it stamps SECRET. Yet the Deputy Secretary of Defense for Counterintelligence and Security conceded under congressional questioning that approximately fifty percent of those classification decisions are overclassifications. Half of an architecture designed to protect the force is protecting nothing—and the friction it generates slows the delivery of everything, including the material that genuinely matters.
The result is an intelligence assembly line that produces enormous volume at enormous cost while failing to deliver synthesis to the people who need it fastest. The problem is not collection. The IC collects more information than any organization in history. The problem is not analysis—brilliant analysts populate every agency. The problem is plumbing. The architecture was designed to move classified material upward through echelons, with synthesis happening at progressively higher levels of command. But in a conflict like Operation Epic Fury, where the threat environment changes hourly across seven domains simultaneously, the people at the bottom of that pyramid need the synthesized picture before the people at the top have finished reading the cable traffic. The architecture delivers too late what it delivers at all.
The Second Track: The Kuwait Proof
Operation Epic Fury provides the real-time proof of concept—not as a hypothetical but as a live demonstration of the information inversion in action. On February 28, 2026, the United States and Israel launched coordinated strikes across Iran under Operations Epic Fury and Roaring Lion. Within forty-eight hours, any analyst with access to open-source reporting—no clearance required, no SCIF needed—could assemble a comprehensive operational picture fusing seven distinct intelligence domains:
Military operations from CENTCOM press releases, IDF statements, and JINSA’s operational updates. Nuclear safeguards from IAEA Director General Grossi’s statement to the Board of Governors on March 2 and subsequent satellite imagery assessments confirming damage at Natanz. Maritime disruption from Kpler’s real-time analysisshowing Strait of Hormuz transits collapsing from twenty-four vessels per day to near zero. Energy markets from Bloomberg, Reuters, and Investing.com, tracking Brent crude surging past ninety dollars per barrel. Diplomatic channels from Reuters, AP, and Al Jazeera, capturing Iran’s Foreign Minister stating there is no reason to negotiate. Cost analysis from CSIS’s estimate that the first hundred hours cost $3.7 billion, roughly $891 million per day, with $3.5 billion unbudgeted. Iranian internal dynamics from Iran International, Fars News Agency, and state media, documenting the Interim Leadership Council, the succession debate, and the Foreign Ministry’s admission that military units have fractured from central control.
No single intelligence directorate within the Department of Defense is chartered to fuse all seven of these streams into a single analytical product and push it to the tactical level in real time. The J-2 handles military intelligence. The J-5 handles policy and strategy. Energy and maritime analysis sits in different shops. IAEA reporting flows through State Department channels. The economic analysis comes from Treasury or specialized commands. Each silo holds genuine expertise. None is chartered to assemble the picture. The result is that a twenty-two-year-old specialist standing post in Kuwait at three in the morning operates on a threat model built from whichever slice of this picture her command decided to brief—while the complete picture is available to anyone with a browser and the training to synthesize it.
Consider what that specialist would know if she had access to the synthesized product. She would know that Iranian retaliatory capability is degrading rapidly in one dimension—ballistic missiles—while increasing in lethality in another—drones. She would know that the Strait of Hormuz closure means the regional economic infrastructure she is stationed to protect is under simultaneous military and economic siege. She would know that Hezbollah has opened a second front in Lebanon, that the IDF has issued evacuation orders covering half a million people in southern Beirut, and that a ground invasion of Lebanon could redirect Israeli military assets away from the Iranian theater.
She would know that Amazon Web Services data centers in Bahrain and the UAE have been knocked offline by drone strikes—meaning the digital infrastructure her unit may rely on for communications and logistics is degraded. She would know that her own government’s stated war aims shifted in the past twenty-four hours from “destroy nuclear capability” to “unconditional surrender”—a shift that changes the timeline, the escalation trajectory, and the likelihood that the conflict she is in will end in weeks rather than months. Every one of these facts shapes her tactical reality. None of them is classified. None of them was in her S-2 brief.
The irony runs deeper. The generation now filling the enlisted ranks grew up synthesizing information across dozens of simultaneous feeds. They are the most information-fluent cohort in military history. The institution responds by handing them a straw and positioning them next to a fire hose—then wondering why the force is surprised when the threat pattern shifts overnight.
The Convergence Gap
The convergence gap is structural, not technological. The technology to fuse open-source streams in real time exists. Commercial platforms do it daily for hedge funds, shipping companies, and news organizations. The gap exists because the defense intelligence architecture was designed during the Cold War to protect against a single monolithic adversary through compartmentation, and it has never been redesigned for an operating environment in which the adversary is a fragmenting regime launching uncoordinated drone swarms across six countries simultaneously.
The 9/11 Commission identified this gap in 2004 when it found that the failure to share information contributed to intelligence gaps before September 11, 2001, and that “the U.S. government did not find a way of pooling intelligence and using it to guide the planning and assignment of responsibilities.” The Commission recommended transforming the intelligence community from a “need to know” system to a “need to share” system. Twenty-two years later, the culture of hoarding has outlived every reform effort. As a Brookings Institution analysis noted, the entire intelligence community was built to follow the Soviet monolith, and the cultural transformation required to address networked, asymmetric threats has been partial at best.
The gap is compounded by what the Brennan Center has called the skewed incentive structure of classification: failure to protect information can end a career, while no one has ever been sanctioned for classifying information unnecessarily. The system defaults to secrecy not because secrecy serves the mission but because secrecy is the path of least personal risk for the classifier. As Supreme Court Justice Potter Stewart wrote in the Pentagon Papers case: “When everything is classified, then nothing is classified, and the system becomes one to be disregarded by the cynical or the careless.” The institution’s own internal culture thus produces the very vulnerability it was designed to prevent.
The Ukraine conflict demonstrated what happens when this gap is partially closed. Open-source analysts tracking Russian force movements, logistics, and casualties through social media, satellite imagery, and electronic intercepts produced strategic-level assessments that rivaled or exceeded classified estimates of Russian defense industrial production. Researchers at the European Journal of International Security found that OSINT-derived models revealed large discrepancies between official Russian claims and actual output—discrepancies that classified channels took months longer to confirm. The lesson was not that OSINT replaces classified intelligence. The lesson was that OSINT synthesis, conducted in real time without compartment walls, consistently delivered faster and often more accurate operational pictures than the stovepiped architecture it was never designed to challenge.
The current conflict makes the Ukraine lesson acute. Iran’s Foreign Ministry admitted on March 3 that its military has lost control of several units operating on prior general instructions. This is not a minor data point. It is a fundamental shift in the threat model for every American soldier in the Persian Gulf. An adversary with centralized command-and-control produces predictable threat patterns. An adversary with fractured command-and-control produces unpredictable, locally initiated actions by units following outdated orders with no oversight. The threat becomes more dangerous precisely because it becomes less coordinated. Any competent tactical analyst given that single piece of information—which was published by Reuters, cited by multiple outlets, and available to anyone with an internet connection—would immediately recognize that the defensive posture briefed forty-eight hours earlier required revision. But the architecture that carries this information to tactical units is not designed for speed. It is designed for control. And control, in this context, is the enemy of survival.
Naming the Weapon
The weapon is The Information Inversion: the structural condition in which the defense classification architecture produces a tactical intelligence environment inferior to what is freely available through open-source synthesis. It is not a bug. It is the predictable output of a system designed to protect secrets from adversaries that simultaneously prevents synthesis across domains, restricts dissemination to echelons that need it most, and incentivizes overclassification at every decision point. The weapon is not wielded by an adversary. It is wielded by the architecture itself. And the people it strikes are not in Washington. They are in Kuwait, at three in the morning, with a threat model that expired six hours ago.
The inversion is most dangerous precisely when it is most invisible. A soldier receiving a classified threat brief has no way of knowing that the brief omits seven-eighths of the operational picture—the maritime disruption data, the energy market signals, the nuclear safeguard status, the diplomatic channel closure, the adversary’s internal fragmentation—because those streams were never fused into the product she received. She cannot miss what she was never shown. The system’s failure is undetectable to the people it fails. They discover the gap only when the threat arrives in a form their brief did not predict—and by then, the discovery is measured in casualties.
The Doctrine
Pillar One: Tactical Fusion Cells. Stand up dedicated open-source fusion cells at the brigade and battalion level, staffed by trained OSINT analysts with the explicit charter to synthesize across military, diplomatic, economic, maritime, and nuclear domains. These cells operate on unclassified systems, produce unclassified products, and push those products to every echelon below them without the friction of classification review. The model exists in embryonic form in the intelligence community’s existing OSINT enterprise. Extend it to the tactical edge where it is needed most.
Pillar Two: The Synthesis Standard. Establish a doctrinal requirement that every threat assessment delivered to forces in contact must include an open-source annex fusing relevant reporting across all available domains—not just the classified take from the unit’s organic intelligence section. The annex is not a supplement. It is a co-equal component of the assessment, produced by the fusion cell, and delivered alongside the classified brief. If the open-source picture contradicts the classified picture, that discrepancy is flagged, not suppressed.
Pillar Three: Classification Accountability. Implement the Brennan Center’s long-standing recommendation for spot audits of classifiers with escalating consequences for serial overclassification. When fifty to ninety percent of classified material does not merit its designation, the system is not protecting the force—it is blinding it. Make the cost of unnecessary classification equal to the cost of unauthorized disclosure. Rebalance the incentive structure so that officers think twice before stamping SECRET on material that belongs on the unclassified net where it can save lives.
Pillar Four: Digital Native Recruitment. Recruit and retain the generation that grew up synthesizing information across simultaneous feeds. Build career tracks that reward OSINT tradecraft, multi-domain synthesis, and real-time analytical production. The twenty-two-year-old specialist who can fuse seven open-source streams into a coherent operational picture in forty minutes is not a liability to be managed. She is the most valuable intelligence asset in the theater. Train her. Equip her. Promote her. Do not bury her behind a system designed for an adversary that dissolved in 1991.
Pillar Five: The Convergence Intelligence Directorate. Establish a permanent Convergence Intelligence Directorate within CENTCOM and each Geographic Combatant Command, chartered specifically to fuse open-source streams across the domains that stovepiped intelligence architectures cannot bridge: military operations, nuclear safeguards, maritime disruption, energy markets, diplomatic signaling, and adversary internal dynamics. This is not a new bureaucracy. It is the institutional recognition that the domains which determine whether soldiers live or die do not respect the organizational chart of the intelligence community—and the force should not have to die while the institution catches up.
The directorate would produce a daily convergence product—modeled on the structure of a comprehensive operational situation report—that fuses all available open-source streams into a single, unclassified analytical document and pushes it to every echelon from combatant command to squad. The product exists to close the gap between what the institution knows and what the force receives. If the concept sounds radical, consider that it is exactly what commercial intelligence firms already do for shipping companies, hedge funds, and insurance underwriters. The defense establishment is the only institution in the world that spends a hundred billion dollars a year on intelligence and cannot deliver a fused operational picture to a specialist standing post.
The Walk
She is twenty-three years old and standing post at Camp Arifjan at 0300. She has been in the Army for fourteen months. She processed more information before breakfast this morning than the entire intelligence staff of a World War II division processed in a week. She does not know that the enemy’s command-and-control architecture fractured overnight, that drone hit rates have quadrupled while missile launches have cratered, or that the threat model she was briefed on twelve hours ago no longer matches the threat she faces tonight. She does not know these things because the classification architecture—built to protect her—has prevented the synthesis that would save her.
Six Americans died in Kuwait in the opening hours of this war. The intelligence existed to understand the threat they faced. The architecture to deliver it to them did not. The information was not hidden by the enemy. It was hidden by the system—buried under fifty million annual classification decisions, half of which the system’s own custodians admit are unnecessary. Chief Warrant Officer 3 Robert M. Marzan, fifty-four, of Sacramento, California. Major Jeffrey R. O’Brien, forty-five, of Indianola, Iowa. Four others whose families were still being notified when their names should have been the last argument anyone needed for tearing down the architecture that failed them.
The intelligence community will respond to this argument with the claim that open-source synthesis cannot replace classified intelligence. That is true. Nobody is claiming otherwise. But the question is not whether OSINT replaces classified material. The question is whether the classification architecture’s inability to deliver synthesized intelligence to the tactical level faster than open-source channels can deliver it represents a structural vulnerability that gets soldiers killed. The answer, measured in the six names from Kuwait, is yes. The architecture that was built to protect the force is blinding it. The information inversion is real, it is measurable, and it is lethal.
The young inherit what the old build. If the architecture blinds the force, the architecture must change. The alternative is to keep handing straws to people standing next to fire hoses and calling it security. The intelligence already exists. The synthesis is possible. The only thing missing is the institutional will to deliver it to the people who need it most—before the next specialist at the next post in the next war becomes the next name on a casualty notification.
The information inversion is the convergence gap. Close it, or count the dead.
RESONANCE
Brennan Center for Justice (2011). Reducing Overclassification Through Accountability. Goitein E, Shapiro DM. https://www.brennancenter.org/our-work/research-reports/reducing-overclassification-through-accountability. Summary: Documents that government officials estimate fifty to ninety percent of classified material does not merit its designation, and proposes accountability mechanisms including spot audits with escalating consequences for serial overclassifiers.
Brennan Center for Justice (2023). The Original Sin Is We Classify Too Much. Goitein E. https://www.brennancenter.org/our-work/analysis-opinion/original-sin-we-classify-too-much. Summary: Argues that the classification system’s skewed incentives—penalties for under-protecting, no penalties for overclassifying—guarantee that busy officials default to secrecy regardless of national security merit. Cites fifty million classification decisions annually.
Center for Public Integrity (2015). Agencies Failed to Share Intelligence on 9/11 Terrorists. https://publicintegrity.org/politics/agencies-failed-to-share-intelligence-on-9-11-terrorists/. Summary: Documents specific instances where FBI, CIA, and other agencies possessed complementary pieces of the 9/11 plot but classification barriers and compartmentation prevented synthesis.
Center for Strategic and International Studies (2026, March 6). Operation Epic Fury Cost Estimate. Cited in Al Jazeera reporting. https://www.aljazeera.com/news/2026/3/6/iran-war-what-is-happening-on-day-seven-of-us-israel-attacks. Summary: Estimates the first one hundred hours of Operation Epic Fury cost $3.7 billion, approximately $891 million per day, with $3.5 billion unbudgeted.
Elwell J, Morrow T (2021). Event Barraging and the Death of Tactical Level Open-Source Intelligence. Military Review, Army University Press. https://www.armyupress.army.mil/Journals/Military-Review/English-Edition-Archives/January-February-2021/Rasak-Open-Source-Intelligence/. Summary: Warns that adversaries will exploit tactical OSINT through “event barraging”—digital inundation with fabricated events—while acknowledging that OSINT at the tactical level provides faster situational awareness than deploying collection assets.
European Journal of International Security (2025). Open Source Intelligence (OSINT) and the Fog of War at the Strategic Level: Defence Industrial Production in Russia. Cambridge University Press. https://doi.org/10.1017/eis.2025.6. Summary: Demonstrates that OSINT-derived models of Russian defense industrial production revealed discrepancies that classified channels took months longer to confirm, establishing OSINT as a viable complement to traditional intelligence at the strategic level.
Hulnick AS (2010). The Dilemma of Open Source Intelligence. In Johnson LK (ed.), The Oxford Handbook of National Security Intelligence. Cited in The Cove, Australian Army. https://cove.army.gov.au/article/tactical-application-open-source-intelligence-osint. Summary: Estimates that eighty percent of the intelligence database is derived from open-source material, establishing OSINT as the foundational layer upon which classified intelligence is built.
International Atomic Energy Agency (2026, March 2). Director General’s Introductory Statement to the Special Session of the Board of Governors. IAEA. https://www.iaea.org/newscenter/statements/iaea-director-generals-introductory-statement-to-the-board-of-governors-2-march-2026. Summary: Grossi reports no radiation elevation above background in bordering countries, confirms IAEA communication with Iran is limited, and warns that a radiological release cannot be ruled out given operational reactors across the region.
JINSA (2026, March 3). Operations Epic Fury and Roaring Lion: Update 1. Jewish Institute for National Security of America. https://jinsa.org/wp-content/uploads/2026/03/Operations-Epic-Fury-and-Roaring-Lion-03-03.pdf. Summary: Documents that Iranian missile campaign rate of fire dropped ninety-five percent while drone hit rate increased from four to twenty-four percent—a shift indicating tactical adaptation that changes the threat model for ground forces.
Kaplan F (2016). Dark Territory: The Secret History of Cyber War. Simon & Schuster. Summary: Documents the intelligence community’s structural inability to share information across agency boundaries, tracing the cultural roots to Cold War compartmentation practices that persist decades after the Soviet threat dissolved.
Kpler (2026, March 1). US-Iran Conflict: Strait of Hormuz Crisis Reshapes Global Oil Markets. https://www.kpler.com/blog/us-iran-conflict-strait-of-hormuz-crisis-reshapes-global-oil-markets. Summary: Reports that the Strait of Hormuz is effectively closed for commercial shipping through insurance withdrawal rather than physical blockade, with limited traffic restricted to Iranian and Chinese-flagged vessels.
Leidos (2025). From Open Source to Operational Insight: How OSINT Is Shaping Modern Intelligence. https://www.leidos.com/insights/open-source-operational-insight-how-osint-shaping-modern-intelligence. Summary: Cites the DIA 2024–2028 OSINT Strategy and the ODNI 2024–2026 OSINT Strategy, both acknowledging that open-source intelligence is now incorporated in nearly all finished intelligence products and that extracting actionable insights from open-source data will only increase in importance.
National Commission on Terrorist Attacks Upon the United States (2004). The 9/11 Commission Report. W.W. Norton. https://www.govinfo.gov/content/pkg/GPO-911REPORT/pdf/GPO-911REPORT.pdf. Summary: Found that “current security requirements nurture overclassification and excessive compartmentation of information among agencies” and recommended transforming the intelligence community from a “need to know” to a “need to share” culture.
NBC News (2023, January 25). America’s System for Handling Classified Documents Is Broken, Say Lawmakers and Former Officials. https://www.nbcnews.com/politics/national-security/americas-system-classified-documents-broken-rcna66106. Summary: Brennan Center expert Elizabeth Goitein states that fifty million classification decisions are made annually, ninety percent of which are probably unnecessary, creating a system impossible to comply with consistently.
Office of the Director of National Intelligence (2024). ODNI Strategy. https://www.govinfo.gov/content/pkg/GOVPUB-PREX28-PURL-gpo234155/pdf/GOVPUB-PREX28-PURL-gpo234155.pdf. Summary: Acknowledges that ODNI is “driving classification reform” while simultaneously noting that the intelligence community must develop structures and mechanisms to promote collaboration across agencies.
Peretti A (2025). The Prometheus Option. CRUCIBEL. Summary: Argues that talent mobility constitutes an asymmetric defense asset and that institutional architecture’s inability to deploy expertise across organizational boundaries represents a strategic vulnerability.
Reducing Over-Classification Act (2010). Public Law 111-258. https://intelligence.senate.gov/laws/reducing-over-classification-act-2010. Summary: Codified the 9/11 Commission’s finding that overclassification and excessive compartmentation nurture intelligence failures, requiring the Secretary of Homeland Security to develop a strategy to prevent overclassification and promote information sharing.
Stremitzer C (2026, February 28). Houthis Signal Renewed Red Sea Shipping Attacks After U.S.–Israeli Strikes on Iran. gCaptain. https://gcaptain.com/houthis-signal-renewed-red-sea-shipping-attacks-after-u-s-israeli-strikes-on-iran/. Summary: Documents that Houthi-controlled Yemen threatened to resume Red Sea attacks following the start of Operation Epic Fury, with BIMCO warning of sharp war risk premium increases if attacks materialize.
U.S. House of Representatives (2007). Hearing on Classification of National Security Information. Committee on the Judiciary. https://www.govinfo.gov/content/pkg/CHRG-110hhrg38190/html/CHRG-110hhrg38190.htm. Summary: Deputy Secretary of Defense Carol A. Haave conceded under questioning that approximately fifty percent of classification decisions are overclassifications. Multiple witnesses testified that Cold War compartmentation culture persists despite the transformation of the threat environment.