The idea of visibility, an “all-seeing-eye,” and, respectively, the efficacy of the image in conveying both knowledge and power, can be traced back to the early days of modern warfare in Europe, and dominates the military imaginary of the twentieth century. More recently, this primacy of the surveillant image for the identification and production of military targets has been called into question by new techniques of data processing and so-called artificial intelligence. 

Broadly speaking, military techniques of imaging may follow one of two functions: firstly, the production of knowledge—the clearing of the “fog of war” as Carl von Clausewitz notoriously put it—which is typically achieved by means of surveillance, what in the German language is ambiguously termed “Aufklärung” [enlightenment], and in English reconnaissance, or not less ambiguously “military intelligence.” When looking at the history of military imaging a second function becomes apparent, namely, the production of discipline or obedience. The ability to produce obedience is necessary for the state to execute its monopoly on violence or, put another way, if a state is not able to produce obedient subjects in its military, it will potentially not be able to defend its territory and lose its sovereignty as a state. That at least is the logic on which modern statehood is grounded, according to Max Weber.¹

The convergence between those two functions, the production of knowledge on the one hand, and the production of obedience on the other, becomes nowhere as apparent as in the context of modern warfare.² The modernity of warfare is itself inscribed in a “culture of violence” that is rooted in a specifically European project: going back to the European Renaissance and the inception of European colonial expansion and leading up to the project of the Enlightenment.

Discipline and Knowledge Production in the Early Modern Period

At the height of the Dutch Military Reforms around the turn of the seventeenth century, the painter and engraver Jacob de Gheyn was commissioned to produce images for The Exercise of Arms, a manual for the development of a new drill technique. The subsequent reforms changed the way war was conducted and, as Antoine Bousquet points out, led the way for the modern concept of a “scientific way of warfare.”³ At the same time the Dutch Republic also rang in the age of global colonialism as a decidedly capitalist project. The Dutch military reform coincides with the beginning of a burgeoning capitalist culture of discovery, military expansion, and economic extraction and accumulation, that serves as a blueprint for subsequent forms of European warfare.

In the late sixteenth century, the Dutch troops under Maurice of Orange were caught in a deadlock against the superior military force of the Spanish Crown. The Dutch commanders responded with sweeping reforms that laid the foundation for modern battlefield tactics. The firearms that were available around 1600—arquebus and musket—were complicated to handle. The Dutch reformers undertook to break down the movements required for the handling of arms into individual segments, each according to a drill command, so that the sequences required for loading, lighting, shooting, and re-loading could be practiced and coordinated. The new regime of training and discipline also allowed the troops to move in a more controlled way across the battlefield, while the arms could now be fired more effectively in synchronised salves.

In de Gheyn’s first commission of a military sujet—The Siege of Geertruidenberg—the artist combined Dutch advances in techniques of mapping with an imagined aerial view of the troops, rendering the battle scene as a kind of projection of the historical events onto the map of the Netherlands. It is safe to say that he was already well-known in the Dutch military circles, when he was commissioned to work on The Exercise of Arms. Here, de Gheyn recorded the correct body postures for the handling of the firearms and pike on elaborate copper plates:

Parallel to his work on the Wapenhandelinghe, de Gheyn also began to work as a draftsman for anatomical drawings at the University of Leyden under the famous botanist and anatomist Pieter Pauw. The methods of analysis, the division into elements, the naming of parts, and the development of a scientific nomenclature, were all hallmarks of the changing relationship between language and image in the field of knowledge around 1600, which evolved from “subjective naming” to “objective designation” in anatomical drawings.⁴ This development is evident, on the one hand, in the attention to detail, the “fixing of the gaze” through the nomenclature that ascribes a word to each individual element—as is evident also in the Wapenhandelinghe’s correspondence between image and command word.⁵

The kind of production and transfer of knowledge that takes place here by means of imaging cannot be fully rendered in language or replaced by words. Thus, a description of the posture in words—written or spoken—would never come close to the visual representation in terms of its clarity or in terms of the speed with which it can be grasped and performed. 

Similarly, describing the minute details of the dissected arm that de Gheyn pictured at University of Leyden in words could never approximate the level of information conveyed by the image. As art historian Svetlana Alpers has shown in a classic study of the so-called “Golden Age” in the Netherlands, Dutch visual culture at the time was marked by a strong appreciation for this specificity of the image as a container for knowledge. 

But adding to that, the particular conditions under which the young Dutch Republic was formed also determined a particularly domineering relationship to land: Not only had the territory to be defended from the Catholic Spaniards, at a different “front,” a large part of the (dry) land of the Dutch Provinces had to be wrung from nature by means of dams and an elaborate system of water drainage. Over 110,000 hectares of land were reclaimed from the sea by dikes and drainage systems, increasing the land mass of North Holland by more than 50 percent.⁶ In the Dutch Republic, more so than in other places, control over the land meant control over nature. Corresponding to this particular attitude to territory, painting, and mapping the landscape became important areas of artistic production. Jacob de Gheyn’s early work on map-like battle-scenes can be seen in this context, but it is also an expression of what Martin Warnke in his book about the Political Landscape has described as a shift towards a vertical view of the battlefield, and what we perhaps would call “drone-like” today: According to Warnke, the commander’s elevated perspective, away from the actual fighting, as if hovering above the battlefield, is an expression of the modern self-image of the military commander. No longer taking part in the battle at ground level, he commands from above, while below “[o]n the battlefield […] the mobile, anonymous, drilled masses fight.”⁷ 

In its dual function as a map and a work of art, de Gheyn’s depiction of The Siege of Geertruidenberg corresponds to the function of taking inventory of dominance that Warnke ascribes to military landscape painting: By projecting the landscape view and map onto each other, the landscape is “memorized, so to speak, the ruler has to know it, to ‘survey’ it, to keep it available.” Accordingly, properties are “to be taken into view” and thus “perceived as a size to be administered, managed, exploited.”⁸

This particular attitude towards landscape reflected the ambitions of the Dutch Confederation on the way to becoming a globalised world power that was (for the first time) a state power founded on capital,⁹ rather than divine right, which set it off from its Spanish opponents. In 1595, Maurice of Orange sent the first Dutch trading fleet to Asia from Amsterdam, which a few years later would transform into the Dutch East India Company. Subsequently, Amsterdam transformed into the largest commodity market and leading capital market in seventeenth century Europe. In 1605, the Dutch East India Company achieved the surrender of the Portuguese fort on the Moluccan island of Ambon in present-day Indonesia, thereby establishing the Dutch Colonial Empire. With art historian Jaleh Mansoor, the Dutch relationship to the depiction of landscape can be described as an expression of these new conditions, in which a “visual militancy”¹⁰ crystallised as a sign of the change from the feudal to the modern economic structures of the sixteenth and seventeenth century. “Wherever they set foot, devastation and depopulation followed,” comments Karl Marx on the colonial exploits of the Dutch as the “model capitalist nation of the 17th century.”¹¹ All Dutch colonial cities were slave societies and the Dutch trade with enslaved people blossomed towards the middle of the seventeenth century, just decades after the publication of the Exercise of Arms. Thus, the visual capture of land and bodies presents, as Ann Jensen Adams puts it, “a visual variation of an economic relation that by the 17th century was firmly established.”¹²

Colonial Continuities and the “Age of the World Target” 

What becomes clear by way of the example of de Gheyn’s drill manual is that the function of acquiring knowledge by means of images easily merges with the function of submitting bodies to military rule. It is useful to keep this in mind when looking at the explosion of imaging techniques at the onset of the twentieth century and beyond. Military imaging makes a significant leap with the fusion of photography and motorised flight, a point that has been discussed at length by Paul Virilio and others:¹³ During the colonial invasion of Libya by the Italians in the year 1911, cameras were for the first time used on board of airplanes, and it was also the first time that bombs were dropped from airplanes onto a civilian population. This first instance of “bombing civilians”¹⁴ from the air marks the beginning of aerial warfare as we know it up to the present day.

With intercontinental nuclear weapons and satellites, military imaging has entered what Rey Chow has called the “Age of the World Target.”¹⁵ Chow asserts that, as there is no place on earth that could not potentially be pictured by military reconnaissance, there is also no place on earth that could not potentially be hit by a nuclear missile. Picturing and targeting have become mutually interchangeable, and the linkage between representation and potential destruction becomes increasingly tight. Chow describes the perspective associated with this “targeted” type of knowledge production as fundamentally self-referential:

As long as knowledge is produced in this self-referential manner, as a circuit of targeting or getting the other that ultimately consolidates the omnipotence and omnipresence of the sovereign “self”/“eye”—the “I”—that is the United States, the other will have no choice but remain just that—a target whose existence justifies only one thing, its destruction by the bomber.¹⁶

When looking at the history of military imaging, it appears that this self-referentiality is the defining feature of knowledge production in the context of warfare: a closed circuit that pictures the enemy, perceived as other, as a potential target, justifying its own destruction, and the “self” as sovereign self, meaning an enlightened self—a surveillant self at war.

Since the turn of the century, unmanned aerial systems generate high-resolution images covering large areas, which are enriched with other data, such as metadata from mobile phone networks and general population data. The drone surveillance takes place mainly outside the US-American territory and often in continuity with former colonial settings. The flood of data-enriched images is increasingly overwhelming for the human eye or for human cognition. Networked algorithms and, more recently, machine-learning applications are now used to automatically interpret images and help with the identification of military targets. The workplace environment of a drone operator has been described by Peter Asaro as the “bureaucratizing of killing.”¹⁷ The automatisation of targeting inevitably means that large parts of the decision-making processes are outsourced to private corporations. The “self-referentiality,” the closed-circuit of target production, is made even tighter when the production of targets follows the logic of those private stakeholder’s interests, rather than of governments and the people they represent.

A recent report by the National Security Commission on Artificial Intelligence was directed by the former CEO of Google, Eric Schmidt. The reason why Google has been so successful as an advertising platform, and not only as a search engine, is the company’s effective way of extracting networked user data that it harvests from their user’s search histories to create very effective targeted ads. Targeted advertising not only paved the way to Google’s enormous financial success, they also laid the cornerstone for the methods of what Shoshana Zuboff has called “surveillance capitalism,” including its “economic imperatives defined by extraction and prediction” of behavior.¹⁸ As Eric Schmidt’s involvement in the National Security Commission on AI-report already suggests, the way targeted ads work can provide insights on the ways that military AI is used for targeting individuals or groups or even whole populations: If it is technically feasible to predict the behavior of individuals and populations, the ”fog of war” is lifted and those individuals and populations become nearly ideal military targets.

In March of 2025 the security company Palantir Technologies delivered the first two prototypes of the “Tactical Intelligence Targeting Access Node“ (TITAN) system to the U.S. Army. Palantir, which was co-founded by Peter Thiel, is known for its security software products, which are based on pattern analysis technologies that were previously developed for PayPal.¹⁹ In Ukraine, Palantir has been embedded in the war effort since mid-2022 and, according to CEO Alex Karp, has been involved in identifying a large proportion of military targets for the Ukrainian military in its attempts to fend off Russian attacks.²⁰ Palantir’s civilian software system Gotham is now used not only by the US police and military and intelligence services around the world, but also for instance by the Hessian police under the name Hessendata.²¹ Gotham collects a wide variety of personal and population data and compiles it on a user-friendly interface in order to visualise “criminogenic” patterns and connections, such as the social environment of criminal suspects, in the form of graphs. 

Targeting today not only means to picture, to represent a potential military target in an image to render it destructible; by operating on the level of data structures within the behavioral data of whole populations, targeting goes way beyond the “world picture,” in the sense of Rey Chow. Some of the technologies that the Israeli Defense Forces use in their assault on the Gaza Strip since October 2023 seem like the realisation of fantasies about the degree of autonomy of “military AI” that have dominated the military discourse so far mainly from a US-American perspective. In April 2024, an investigative report by Yuval Abraham revealed that the IDF used an AI-supported system called Lavender for the identification of many of their targets for aerial bombardment. The magazine Foreign Policy described the system as a “mass assassination program of unprecedented size.”²²

The Lavender software analyzes information collected on most of the 2.3 million residents of the Gaza Strip through a system of mass surveillance, then assesses and ranks the likelihood that each particular person is active in the military wing of Hamas or PIJ. According to sources, the machine gives almost every single person in Gaza a rating from 1 to 100, expressing how likely it is that they are a militant.²³

Lavender apparently relies on large-scale surveillance of population data, in which machine learning applications are used to identify anomalies (such as frequent SIM card changes or any contact with members of militant groups) and visualise them as graphs. As the investigative research claims, the results produced by Lavender were treated as “orders” to be followed without question.²⁴ The use of an advanced system of military AI for essentially targeting an enclosed population as a whole, a population that has absolutely nowhere to go, presents a situation in which the closed circuit of military targeting has become complete. The knowledge produced through the extensive, if not total, surveillance of a population will never reveal anything other than their legitimacy as a target. This is the self-referentiality, to which Rey Chow has referred as a “circuit of targeting […] that ultimately consolidates the omnipotence and omnipresence of the sovereign ‘self’” and renders the other as “a target whose existence justifies only one thing, its destruction by the bomber.”²⁵

Footnotes

1. Max Weber, Wirtschaft und Gesellschaft. Grundriß der verstehenden Soziologie, Tübingen: Mohr Siebeck, 1980, p. 822. For a detailed history of military obedience from a European perspective, see: Ulrich Bröckling, Disziplin. Soziologie und Geschichte militärischer Gehorsamsproduktion, München: Fink, 1997. ↑

2. In this sense, warfare forms a gap in the analysis of this convergence of power and knowledge within the work of Michel Foucault, a gap that he addressed and repeatedly promised to fill in his 1976 lectures at the Collège de France under the title Society Must Be Defended (New York: Picador, 2003). ↑

3. See: Antoine J. Bousquet, The Scientific Way of Warfare. Order and Chaos on the Battlefields of Modernity, New York: Columbia University Press, 2009, p. 57; see also: Ulrich Bröckling, Disziplin. Soziologie und Geschichte militärischer Gehorsamsproduktion, Munich: Fink, 1997, p. 31. ↑

4. See: Marielene Putscher, Geschichte der medizinischen Abbildung. Von 1600 bis zur Gegenwart, Munich: Heinz Moos Verlag, 1972. ↑

5. Ibid., pp. 15–16. ↑

6. Ann Jensen Adams, “Competing Communities in the ‘Great Bog of Europe,’” in  Landscape and Power, ed. W. J. T. Mitchell, Chicago: University of Chicago Press, 1994, p. 41. ↑

7. Martin Warnke, Politische Landschaft. Zur Kunstgeschichte der Natur, Munich: Hanser, 1992, p. 68 (author’s own translation). ↑

8. Ibid., p. 65 (author’s own translation). ↑

9. Jaleh Mansoor, “Militant Landscape. Notes on Counter-Figuration from Early Modern Genre Formation to Contemporary Practices, or, Landscape After the Failure of Representation,” ARTMargins, vol. 10, no. 1 (2021): pp. 20–38. ↑

10. Mansoor, “Militant Landscape,” p. 20. ↑

11. Karl Marx. Capital, vol. 1., trans. Ben Fowkes. London: Penguin Books, 1990, p. 916. ↑

12. Adams, “Competing Communities in the ‘Great Bog of Europe,’” p. 58. ↑

13. Paul Virilio, War and Cinema: The Logistics of Perception, London: Verso, 1989; Friedrich Kittler, Grammophon, Film, Typewriter, Berlin: Brinkmann & Bose, 1986, pp. 177–203; Bernhard Siegert, “Luftwaffe Fotografie. Luftkrieg als Bildverarbeitungssystem 1911 + 1921,” Fotogeschichte, Jg. 12, 45/46, 1992, pp. 41–54. ↑

14. See: Yuki Tanaka, Marilyn B. Young (eds.), Bombing Civilians. A Twentieth-Century History, London: The New Press, 2009. ↑

15. Chow here refers to the somewhat dubious concept of the “world picture” that Martin Heidegger developed in 1933 and, as Sidonie Kellerer has shown, used as a cover for his active involvement with the Nazis. See: Sidonie Kellerer, “Heideggers Maske. ‘Die Zeit des Weltbildes’ – Metamorphose eines Textes,” Zeitschrift für Ideengeschichte, vol. 5, no. 2 (2011): pp. 109–120. See also: Martin Heidegger, “Zeit des Weltbildes,” in Gesamtausgabe, Band 5, Holzwege, ed. Friedrich-Wilhelm von Herrmann, Frankfurt: Vittorio Klostermann, 1977, pp. 75–114. ↑

16. Rey Chow, The Age of the World Target, Durham: Duke University Press, 2006, p. 41. ↑

17. Peter M. Asaro, “The Labor of Surveillance and Bureaucratized Killing. New Subjectivities of Military Drone Operators,” Social Semiotics 23.2, 2013, pp. 196–224. ↑

18. Shoshana Zuboff, The Age of Surveillance Capitalism, eBook, p. 125, 127. ↑

19. Paige Leskin and Rosali Chan, “Palantir, a secretive tech company started by members of the ‘PayPal mafia’ with close ties to the Trump administration, could be one of the biggest tech IPOs ever. Take a closer look at how it makes money,” Business Insider, July 7, 2020, https://www.businessinsider.com/palantir-ice-explainer-data-startup-2019-7?op=1, accessed January 9, 2026. ↑

20. Vera Bergengruen, “How Tech Giants Turned Ukraine Into an AI-War Lab,” Time Magazine (2024), https://time.com/6691662/ai-ukraine-war-palantir/, accessed January 9, 2026. ↑

21. Leonie Sontheimer, Lisa Hegemann, and Gregor Becker, “Palantir Technologies. Die geheimnisvollen Datensortierer,” ZEIT Online, September 30, 2020, https://www.zeit.de/digital/internet/2020-09/palantir-technologies-daten-analyse-boersengang-peter-thiel-alex-karp/komplettansicht, accessed January 9, 2026. ↑

22. Simon Frankel Pratt, “When AI Decides Who Lives and Dies. The Israeli military’s algorithmic targeting has created dangerous new precedents,” Foreign Policy (2024), https://foreignpolicy.com/2024/05/02/israel-military-artificial-intelligence-targeting-hamas-gaza-deaths-lavender/, accessed January 9, 2026. ↑

23. Yuval Abraham, “‘Lavender’: The AI Machine Directing Israel’s Bombing Spree in Gaza,” +972 Magazine (2024), https://www.972mag.com/lavender-ai-israeli-army-gaza/, accessed January 9, 2026; see also: Bethan McKernan and Harry Davies, „‚The machine did it coldly’: Israel used AI to identify 37,000 Hamas targets,” The Guardian, April 3, 2024, https://www.theguardian.com/world/2024/apr/03/israel-gaza-ai-database-hamas-airstrikes, accessed January 9, 2026. ↑

24. Abraham, “Lavender.” ↑

25. Chow, The Age of the World Target, p. 41. ↑