What OSINT can and cannot tell you about geopolitical risk

Why more data does not mean better intelligence, and what aviation decision-makers actually need

The volume of publicly available threat data has never been higher. And yet aviation decision-makers are not better protected than they were a decade ago. The problem is not access to data. It is the gap between raw data and verified intelligence.

What OSINT does well

Open-source intelligence has genuine value. ADS-B tracking provides real-time visibility of aircraft movements. Satellite imagery detects activity changes at military installations. GNSS monitoring identifies jamming and spoofing patterns that frequently precede military operations.

A 2022 joint paper by RUSI and the Alan Turing Institute found that publicly available information now forms a majority of the information used to develop actionable national security insights for UK government decision-makers. Open-source methods are indispensable. The question is understanding where their value ends.

Source: RUSI and Alan Turing Institute, The Future of Open-Source Intelligence for UK National Security, June 2022

The structural limits of open source

Relying on open source intelligence alone creates predictable blind spots. The limitations are not technical; they are inherent to the method itself. Four structural weaknesses define where open source fails:

Circular reporting.

During any high-profile escalation, a single source propagates across dozens of outlets, generating the appearance of corroboration where none exists. Decision-makers presented with ten sources reporting the same threat may reasonably infer credibility. If all ten derive from one upstream input, that inference is wrong. The Iraq weapons of mass destruction failure remains the most consequential historical example of this dynamic.

Adversarial manipulation.

State actors can deliberately seed open-source channels with misleading information: inflating threat signals to trigger operational disruption, or suppressing indicators to keep airspace open when it should not be. OSINT has no reliable mechanism to distinguish authentic signals from those designed to deceive.

The streetlight effect.

Analysis naturally clusters around the data that is most accessible, not necessarily where risk is highest. Conflict zones with limited digital infrastructure or state-controlled information environments are precisely the locations where open-source coverage is thinnest and consequences are most severe.

NOTAM lag.

State-issued advisories carry the same structural limitation. The Dutch Safety Board noted that states involved in armed conflicts rarely proactively close their airspace, even where sufficient grounds exist. The gap between what states know and what they publish is structural, not administrative.

Between 14 and 17 July 2014, 61 operators from 32 countries routed their flights through this airspace. On the day of the crash, 160 flights flew over the area before the airspace was closed. The question that arose was: why were those aircraft flying over the eastern part of Ukraine?

Source: Dutch Safety Board Final Report, October 2015

61 operators from 32 countries flying over eastern Ukraine before MH17

12+ FIRs closed simultaneously in February 2026

2,280+ flights cancelled within 24 hours

The Osprey difference

Every signal entering the Osprey platform is assessed against four criteria before it informs a forecast or triggers an alert. This is what separates verified intelligence from open-source aggregation.

Is it independently corroborated?

Osprey monitors more than 200,000 sources across 60 languages. Each signal is cross-referenced against independent sources before it is treated as verified. A story appearing in 40 outlets from a single origin counts as one data point, not 40.

Does it represent a genuine pattern change?

Machine learning models identify deviations from baseline: unusual military movements, shifts in AIS transponder behaviour, communications blackouts, and changes in light intensity in areas of known tension. These are the signals that do not appear in news feeds.

What does the silence mean?

In a world of constant information, what speaks loudest is silence. When the volume of activity around a specific location drops sharply, that absence is itself an intelligence signal. Osprey monitors for when things go dark.

What is the probability and timeframe?

Every Osprey forecast includes a probability assessment and an analyst-reviewed timeframe, not a binary safe or unsafe judgement. In the first half of 2025, 43 percent of Osprey notifications were proactive, with a 98 percent accuracy rating.

200,000+ sources monitored

98% forecast accuracy

60 languages monitored

Source: Osprey Forecasts information sheet; Osprey platform data

The intelligence standard aviation requires

ICAO Doc 10084 sets out the expectation that operators conduct their own independent risk assessments rather than defer to state-issued advisories. That principle creates a clear operational requirement: verified, confidence-rated intelligence delivered with enough lead time to implement route changes as planned decisions rather than reactive cancellations.

The operators best positioned to respond to the 2026 Geopolitical Crisis in the Middle East were not those with access to more data. They were those with access to better analysis.

Speak to our team

Open-source intelligence surfaces the signals. Osprey determines which ones are real.

Sources