Open source surveys – based on the collection of information available on the Internet – are increasingly used by the news media as well as by propaganda communication. They raise many ethical and methodological questions.

On April 5, 2024, Le Monde published a very detailed investigation , showing, through multiple sources accessible online, the destruction of Palestinian hospitals, schools and universities caused by the bombings of the State of Israel in Gaza.

These open source investigations, also called OSINT for “open source intelligence”, are now a recognizable visual and media component of our daily lives. From a collection of information freely available on the Internet (sometimes compared to others drawn from a “field” experience, on site), these elements (photos and videos present on social media, digital maps, administrative documents, government data, scientific articles, etc.) now aim to document a wide range of events.

How and who uses these sources?

Many professionals, but also ordinary citizens, are launching themselves into these investigations at high speed, relying on increasingly sophisticated methodologies and using the power of collective intelligence and the networking of information through social media.

Some independent collectives, professional or not, are formed like Forensic Architecture and Bellingcat and have become, in recent years, essential references for this type of practice. Today, major press organizations like Le Monde , the BBC , or the New York Times have built teams specially dedicated to digital investigation. Others like Agence France-Presse (AFP) or the Global Investigative Journalism Network freely provide resources to train in these techniques.

However, based on a digital surveillance system already widely worked on by academics, such as Shoshana Zuboff for example, and associations such as Quadrature du Net , OSINT raises a series of ethical questions.

The example of al-Ahli hospital

On October 17, 2023, al-Ahli Hospital in the Gaza Strip was hit by a missile. Common sense leads some media to consider that if a missile hits a building in a conflict zone, it is probably the responsibility of the opposing belligerent. In fact, Palestinian Authority President Mahmoud Abbas and Hamas blame the Israeli state for the attack .

However, the Israeli army spokesman announced in a press conference that the hospital explosion was the result of a missile coming from inside the Gaza Strip. To support its statement, the Defense Ministry produced a set of visual documents resembling OSINT and supposed to prove the non-responsibility of the Israeli state in this attack. Press organizations such as the New York Times , Channel 4 or associations such as Forensic Architecture, expressed their doubts about the version put forward by the Israeli army spokesman.

Other media outlets, including Le Monde and Bellingcat , have expressed doubts about the results put forward by their colleagues. They question the very possibility, given the few sources available at the time, of being able to determine those responsible for this explosion.

In fact, some of the investigations into this event are based on partial sources from which it is difficult to draw reliable causal links. In general, when OSINT investigations are conducted by several news outlets or NGOs, the different investigators arrive at the same results, because the methodology is based on the possibility, given to everyone, to redo the investigations from the available sources. However, this is not the case here.

Shortly after, some analyses, such as that of the New York Times , apologized for having potentially poorly covered the event and qualified their version . Others, on the contrary, persevered, like Forensic Architecture .

Our role as academics is not necessarily to hand out good or bad points on the conduct of OSINT investigations, but this case seems relevant to keep in mind, because it corresponds to a first media controversy where OSINT is used by press organs whose words carry into the public debate, without the latter agreeing on the facts.

Another case: Boutcha

We have already noticed a controversy of a similar kind regarding investigations into the Buchá massacre in Ukraine, a series of war crimes committed near kyiv between February 27 and March 31, 2022 during the occupation of the city by the Russian army.

The open-source investigations into the narrative surrounding this tragedy were provided on one hand by a world-renowned newspaper, the New York Times , and on the other hand by pro-Russian propaganda videos circulating on Telegram channels.

In this asymmetry regarding the authority to disseminate information, both sides, with different formats and interpretations, have circulated videos reusing the codes of OSINT investigations.

For example, based on a video taken by Ukrainian soldiers, the New York Times indicates corpses on the ground. Pro-Russian propaganda uses enlargements of these same images to “show”, on the contrary, that we see a corpse getting up in a rearview mirror (in fact a drop of rainwater distorting the image) or that the hands of the bodies are not those of dead people.

“Polysemic footprints”

Theorist Aurélie Ledoux has attempted to show that, in the case of Boutcha, this opposition between traditional media and propaganda stemmed from a counter-discursive logic of open-source investigations. That is to say, from a certain emphasis on OSINT as a counter-discourse that undoes official discourses and that can be easily recovered by conspiracy discourses and disinformation.

She also highlights how information in OSINT investigations is a “polysemic fingerprint” that can embrace different interpretations. Indeed, in both the New York Times ’ work and the pro-Russian videos, it is sometimes difficult to see in the images what the video’s authors are telling us. Thus, the viewer is sometimes faced with enlarged satellite photographs in which he can only distinguish one pixel darker than the others, but in which he is supposed to distinguish a corpse or a vehicle, for example.

The reception we have of these videos depends on the trust we attribute to a particular media or, on the contrary, on a form of distrust towards authoritative discourses which will make some people prefer “alternative” versions and “toxic narratives”, to use an expression of the Italian theorist Wu Ming .

The Visual Grammar of Veridiction

The use of open-source investigative techniques has given journalists the ability to challenge official narratives and unearth stories they might not have otherwise revealed.

However, it can be seen that this approach has also fueled disinformation by giving a false impression of certainty to claims based on objective and often unvalidated documents. It seems to us that this false impression of certainty is based on the use of a certain visual grammar claiming to tell the truth – which we have called “visual grammar of veridiction”.

Indeed, video transcripts of OSINT investigations often set up a very particular time flow. Starting from an initial situation (for example an explosion at time T), they go back in time and reconstruct the events point by point in order to bring us to a final situation (an assertion as to the responsibility for this explosion).

Each step is developed, as for a scientific demonstration, by logical leap from one proof to another, the viewer following the course of scientific reasoning, without necessarily perceiving that it is a cut on the reality of the events that occurred at that moment and in that place. This video reconstruction is based on a set of visual forms constituting a demonstration. But these are empty demonstrative forms allowing a reuse in contexts where the search for truth is not the main concern.

These visual forms are composed of a large but limited repertoire circulating between media. We can note those that are most frequently used: the juxtapositions of satellite images and ground views with colored geometric shapes creating correspondences between the two; the implementation of timelines referring to the aesthetics of video editing software interfaces; the massive use of 3D modeling to restore events of which images are available or not; the presentation of documents as if to testify to the authenticity of the sources used; or the creation of panoramic images by assembling different shots.

A competition for the framing of reality

This visual grammar raises questions when used by the journalistic community. OSINT videos are based on images that record a certain number of signs testifying to a certain relationship with reality. They thus inherit a whole tradition considering that images are proof, yet deconstructed by the academic work of Vincent Lavoie or Aurélie Ledoux and Dork Zabunyan, among others .

As with Boutcha, by focusing on one detail of the images rather than another, and by linking them with different details of other images or information, alternative narratives about the events are put in place, transforming themselves into a staging, despite the other elements of analysis.

A competition for the framing of reality occurs, and OSINT investigators must operate on the understanding that the world, images, and information they use are open to different interpretations, but that factual truth exists. It is a ridge they walk and are still taming in the rendering of their open-source investigations.

They must also tame a second one regarding the dissemination of the results of their investigation. There is indeed a balance to be maintained between the logical and narrative sequence of their demonstrations (on which the attention of the spectator depends as well as the effectiveness of the speech) and the unveiling to the public of the “black box” of tools and choices that have made it possible to construct an interpretation of the events as rigorous as possible.

Author Bios: Allan Deneuville is a Lecturer in Information and Communication Sciences at Bordeaux Montaigne University and Jacopo Rasmi is a Lecturer at Jean Monnet University, Saint-Étienn