Open Source Intelligence

Open Source Intelligence, or OSINT for short, roughly means “reconnaissance from open sources.” What is meant is information that anyone can legally view: newspaper articles, videos, photos on social networks, company registers, flight data, maps, or satellite imagery. “Open source” here refers to freely accessible sources – not disclosed program code, for which the same term is otherwise often used. The trick lies not in obtaining individual facts, but in piecing them together. Anyone who skillfully links a hundred public snippets can derive from them a picture that no single source shows. That is why OSINT counts among the working methods of intelligence services, journalists, investigators, and companies alike.

Why open sources reveal more today than they used to

Thirty years ago, information was scarce. Anyone who wanted to know what was happening in a particular place needed someone on the ground. Today, people voluntarily upload billions of photos, videos, and location data to the internet. As a result, the gap between secret and public knowledge has shrunk.

A well-known example is the research group Bellingcat. Using publicly accessible photos, videos, and satellite images, it reconstructed where the missile came from that shot down flight MH17 over Ukraine in 2014. No stolen documents were needed, just patient comparison of images, shadows, street signs, and timestamps. Such research was later used in court.

OSINT is also relevant to the financial world. Analysts count cars in supermarket parking lots using satellite images to estimate sales figures. Companies use public registers to check whether a business partner is under sanctions. At the same time, attackers use the same methods: anyone who knows the names, positions, and vacation photos of employees can write more convincing fraud emails.

From individual finding to reliable statement

OSINT usually proceeds in four steps. First, a clear question is defined, for example: Where was this video recorded? Then material is collected from as many sources as possible. Next, it is checked and cross-referenced. In the end, there is a statement along with an honest indication of how certain it is.

Two techniques come up constantly in this process. Geolocation means determining where an image was taken – using mountain silhouettes, billboards, power poles, or road layouts that are compared with map services. Chronolocation determines the time, for instance via the cast of shadows, the weather on that day, or the state of a construction site in the background. Both are painstaking work and often take hours for a single video.

This is where AI comes in. Language models summarize large amounts of text, image recognition sorts thousands of photos by subject, and translation systems open up foreign-language sources. This mainly speeds up the collection process. Evaluation remains tricky, because AI systems occasionally invent plausible but false details. Serious OSINT work therefore requires that every claim be traceable back to a verifiable source.

Who uses OSINT – and where the line is drawn

In the news, you mostly encounter OSINT in the context of wars and crises. When a newsroom writes that a video has been “verified,” this method is almost always behind it. Fact-checks of alleged disaster images are also based on it: often it turns out that a photo is years old and comes from a different country.

Beyond that, there is a growing market. Security firms sell OSINT tools to banks, insurers, and government agencies. HR departments look at applicants' public profiles. And in IT security, companies check what information about themselves is freely available online before someone else does.

A common misconception is: whatever is public may be used however one likes. Legally, that is not true. In Germany and the EU, the General Data Protection Regulation protects personal data even when it is freely accessible. Massively combining public information into a personal profile can be prohibited. The line between research and surveillance is thin – and that is exactly what is currently being debated in politics.