What Makes Trevor Paglen’s Work So Powerful?

Trevor Paglen’s work is powerful because he reveals that the contemporary world is shaped by images, infrastructures, and systems that most people rarely see—and that increasingly do not need human beings in order to operate.

A distant military base shimmers through atmospheric distortion. A surveillance satellite crosses a field of stars. A drone appears as a nearly invisible speck against the sky. Fiber-optic cables lie beneath the ocean carrying vast quantities of data. A grid of photographs shows the training images through which computer-vision systems learn to classify human beings. Clouds and landscapes are processed by algorithms whose internal categories are entirely invisible to the viewer.

A photograph may look beautiful, atmospheric, and almost romantic, while documenting machinery designed for surveillance, targeting, data extraction, or secrecy.

Paglen’s central inquiry might be stated this way:

How can art make visible the hidden systems that increasingly determine what can be seen, known, classified, predicted, and acted upon?

His work does not simply expose secret objects. It investigates the deeper politics of perception: Who is allowed to see? Who is being seen? What remains hidden? What happens when machines begin interpreting images for other machines rather than for people? And what kind of world emerges when vision becomes infrastructure?

What the Viewer Encounters

Paglen works across photography, sculpture, moving image, research, writing, engineering, geography, installation, and collaborative technological projects. His practice has examined covert military installations, spy satellites, drone warfare, undersea communications cables, classified government programs, artificial intelligence, facial recognition, influence operations, and the automated interpretation of images. (Pace Gallery)

The viewer may encounter:

• blurred photographs of distant military sites;
• night skies crossed by satellite trails;
• images of drones against clouds or sunlight;
• maps and documents revealing hidden infrastructures;
• photographs of beaches where undersea cables come ashore;
• machine-learning datasets;
• computer-generated image classifications;
• algorithmically altered landscapes;
• sculptural satellites that serve no practical military or commercial purpose;
• video feeds that resemble operational surveillance interfaces.

Many of these works are visually restrained. A photograph may initially resemble a landscape, an astronomical image, a cloud study, a sunset, an abstract color field, or a minimalist grid. Only later does the viewer learn that the small light in the sky is a surveillance platform, that the distant architecture belongs to a classified facility, or that the image has been filtered through a machine-vision system.

The work operates through delayed recognition. Beauty draws the viewer closer; information changes what beauty means.

The Inquiry Beneath the Work

Modern power does not always appear through visible monuments, uniforms, or public institutions. It often operates through hidden or ordinary infrastructures: satellites, databases, cables, classification systems, server farms, algorithms, sensors, proprietary platforms, automated weapons, and image-recognition systems. These structures shape political and social life while remaining largely inaccessible to public observation.

Paglen’s work asks: How can democratic societies understand systems whose operations are secret, technically obscure, physically remote, or deliberately concealed?

This is not only a problem of hidden information; it is a problem of perception. A secret base may exist in a desert, but be located so far away that conventional vision cannot reach it. A satellite may pass overhead, but remain invisible without specialized knowledge and equipment. A facial-recognition model may classify a person, but the categories through which it operates remain hidden in code, training data, and proprietary systems. The contemporary world therefore contains many things that are physically real but perceptually unavailable. Paglen’s artistic project begins in that gap.

Making the Invisible Visible

Paglen is frequently described as an artist who makes invisible systems visible. That phrase is useful, but incomplete. The work does not usually reveal hidden structures with total clarity. His photographs often remain blurred, distant, incomplete, or ambiguous. A secret installation may appear only as a heat-distorted cluster on the horizon. A satellite may be represented by a thin streak crossing the sky. A drone may be barely distinguishable from dust on the lens.

This limited visibility is essential. The work does not say, “Now the secret has been completely exposed.” It says, “Here is evidence that something consequential exists beyond ordinary access.” The image makes the limit of seeing visible.

Paglen’s photographs frequently depend on extreme distance, specialized lenses, astronomical calculations, geographical research, and sustained waiting. The resulting distortions are not technical failures; they show how secrecy is produced spatially and technologically. (The Guardian) The work reveals not only the hidden object, but the structural difficulty of reaching it.

The Landscape of Secrecy

Paglen often describes his practice through the history of landscape art. His landscapes, however, contain infrastructures that conventional landscape imagery tends to exclude: classified bases, weapons-testing zones, intelligence facilities, data cables, satellites, drones, radar systems, and server networks. This changes what landscape means.

A landscape is not simply nature observed from a distance; it is territory organized by power. The desert may contain an invisible military geography, the ocean floor may contain communications infrastructure, and the sky may contain thousands of surveillance and tracking objects. The cloud becomes both an atmospheric phenomenon and a metaphor for data infrastructure. Paglen’s landscapes show that no apparently empty space is neutral. The landscape has become crowded with systems whose visibility is unevenly distributed.

Distance as Political Form

Distance plays a major role in Paglen’s work. A military installation may be hidden because it is geographically remote, legally restricted, surrounded by controlled land, excluded from ordinary maps, or protected by secrecy laws. Paglen often photographs such sites from publicly accessible locations miles away.

The distance becomes visible in the image. Heat haze breaks the architecture apart, light scatters, and forms lose precision. The photograph carries the political relationship between observer and subject:

• The state possesses: proximity, access, technical clarity, and operational knowledge.
• The public possesses: distance, partial evidence, speculation, and blurred visibility.

The image is therefore not merely a photograph of a secret facility; it is a photograph of unequal access to reality.

The Technological Sublime

Many of Paglen’s photographs are visually beautiful, connecting the work to the history of the sublime—the aesthetic experience of forces that exceed ordinary human scale or comprehension. Historically, the sublime was associated with storms, mountains, oceans, and cosmic space. Paglen relocates the sublime into technological and political systems.

The overwhelming force is no longer only nature; it is also surveillance, planetary computation, military technology, and artificial intelligence networks. A night sky remains beautiful, but it is crossed by spy networks. A cloud remains atmospheric, but may be read by computer vision. A desert remains vast, but contains weapons systems and classified architecture. The technological sublime is attractive and threatening at once.

Spy Satellites and the Reversal of the Gaze

Paglen has spent years identifying and photographing classified or unacknowledged satellites by combining astronomical observation, orbital data, research, and collaborative networks of amateur observers. His satellite photographs make visible objects designed to observe the Earth while remaining difficult for the public to identify. (The Guardian)

This produces a reversal: the satellite looks down, and Paglen looks back. The object designed to observe becomes the object observed. But the relationship is not equal. The satellite possesses high-resolution sensors, governmental backing, and massive operational capacity. The artist possesses a telescope, mathematical calculations, and public visibility. The act of looking back does not neutralize the surveillance system; it creates symbolic and political resistance. To observe the observer is to refuse complete transparency.

The Sky Is No Longer Neutral

For most of art history, the sky has functioned as atmosphere, weather, divine realm, or a symbol of transcendence. Paglen shows that the sky is also infrastructure, filled with reconnaissance satellites, communications systems, navigation networks, and experimental tracking objects. The night sky becomes politically occupied.

The technological systems governing life often hide inside perceptual backgrounds that people have learned to ignore.

The work retrains attention: a speck becomes a drone platform, a streak becomes a classified orbit, and a cloud becomes a machine-readable object.

Drones and the Asymmetry of Vision

Paglen’s drone photographs often show aircraft at extreme distance, appearing tiny, indistinct, or almost absorbed into the blue sky. The drone’s visual insignificance contrasts sharply with its operational power. A machine that appears as a speck to the human observer may see the ground below with extraordinary precision. This creates an asymmetry: the drone can see clearly while remaining nearly unseen.

That asymmetry is central to contemporary power systems. Those controlling the machine remain physically distant from the consequences of its actions, while those being observed may not know when they are visible, how they are classified, or who is making automated decisions about them. Paglen’s photographs do not show what the drone sees; they show how little the observed world can see of the drone.

Operational and Invisible Images

Some images are made for people to look at, while others are made in order to perform tasks. A targeting image helps guide a weapon, a facial-recognition image identifies a person, and a machine-vision system interprets traffic data. These are operational images because their primary purpose is action rather than contemplation. Photography is no longer only a human-to-human exchange; it is a loop where a sensor captures data, a machine classifies the image, and a system acts.

Paglen uses the phrase invisible images to describe images produced by machines for interpretation by other machines. These images circulate at an enormous scale without ever becoming visible in ordinary public culture. Yet the political stakes are significant, influencing policing, border control, employment screening, credit scores, and military targeting. Representation becomes power without becoming public appearance.

A Study of Invisible Images and Datasets

In A Study of Invisible Images, Paglen examined the image systems underlying artificial intelligence, including machine-training datasets, computer-vision classifications, and generated images. (Smithsonian American Art Museum)

Grids of training images may appear visually neutral, but each image has been assigned labels. A human face may be classified according to age, emotion, gender, race, attractiveness, or behavioral expectation. These categories are not simply discovered inside the image; they are imposed through social assumptions, dataset construction, and institutional goals. The machine learns not only how the world looks, but how someone decided the world should be divided. Classification is never neutral: a dataset does not merely describe the world, it creates a framework through which the world will be acted upon.

Training Humans

Paglen’s collaborations with researcher Kate Crawford examined how artificial-intelligence systems are trained through enormous datasets and how those datasets reproduce historical assumptions about race, gender, labor, and social value. The title Training Humans reverses a common formulation. Humans train machines, but once automated categories enter institutions, people adapt to them.

A person learns how to pose for facial recognition, how to avoid triggering automated moderation, how to write for search engines, and how to make themselves legible to platforms. The classification system begins shaping the subject it claims merely to observe. A character constructs a performance not only for another person, but for a machine. Identity becomes algorithmically mediated.

Ethical Pressures Inside the Critique

Paglen and Crawford’s treatment of facial-image datasets has also generated critique. Academic creators of specific datasets disputed aspects of their account, arguing that their exhibition and writing mishandled the dataset’s historical context and raised unresolved questions about image consent. (arXiv)

This criticism exposes a contradiction within research-based art: an artwork may critique surveillance, extraction, and the nonconsensual use of images while still relying upon images of people whose participation remains ethically complicated. The critique can accidentally reproduce part of the condition it opposes. This makes the ethical standard more demanding, showing that artists investigating image power must also examine their own power to collect, display, interpret, and recirculate images.

Adversarially Evolved Hallucinations

Computer-vision systems are often described as “seeing,” but machine vision does not resemble human perception. A person sees a face, a gesture, or a social situation; a machine processes features, features vectors, edges, and confidence scores. Paglen’s work explores the mismatch between these systems.

Through Adversarially Evolved Hallucinations, Paglen created images designed through machine-learning loops to activate specific classifications within computer-vision systems. These images may look abstract, strange, or dreamlike to human viewers while being recognized confidently by a machine as a specific object or category. The work exposes a critical distinction:

Recognition is not the same as understanding.

A machine can classify an image confidently while being wrong in ways that appear absurd to people. Yet these confidence scores can still influence real decisions, turning the machine’s hallucination into an institutional reality.

Clouds Seen by Machines

Paglen’s Cloud works transform skies and landscapes through computer-vision operations and algorithmic processing, marked by exhibitions like A Color Notation. (Pace Gallery) Clouds are especially important because they sit between form and formlessness. Human beings recognize images in clouds through psychological projection; machines divide clouds into segments, lines, and statistical features.

The cloud becomes a test of how perception imposes structure upon ambiguity. A person sees a face or animal; an algorithm sees edges and regions. Neither receives the cloud neutrally; each system constructs a world from it.

The Problem of Proof and Research

A documentary photograph is often expected to prove something. Paglen’s images frequently provide evidence, but not definitive proof. A distant blur may be identified through extensive research, but the photograph alone does not establish what the viewer is seeing. The image depends entirely upon captions, coordinates, titles, maps, technical knowledge, and institutional research.

This creates a productive tension. The work critiques secret systems while asking viewers to accept information they cannot independently verify. The viewer remains caught between skepticism and belief, mirroring the world of secrecy itself, where citizens know hidden systems only through leaks, fragments, and indirect evidence. The photograph becomes the visible threshold of a larger inquiry, expanding photography from simple image-making into a method for investigating the physical organization of reality.

Geography as Artistic Method

Paglen trained in geography as well as art, and his work frequently treats geography as the study of how power becomes spatial, tracking orbital paths, restricted zones, oceanic cable routes, and global sensing networks. (Kinoki) The work asks not only what the system is, but where it is. Paglen reveals that digital power still requires physical geography: the cloud has cables, the algorithm has servers, the satellite has an orbit, and supposedly immaterial networks must touch the Earth.

Undersea Cables and the Material Internet

The internet is often imagined as immaterial, moving effortlessly through a metaphorical cloud. Paglen’s work on undersea fiber-optic cables reveals the material infrastructure beneath that metaphor. Data travels through physical cables, landing stations, switches, and server farms across physical jurisdictions. These cables carry enormous volumes of global communication and become primary sites of state surveillance and strategic control. The ordinary landscape contains invisible circulation; a beach used for recreation is also a point through which planetary financial, military, and personal data passes.

Autonomy Cube

Paglen collaborated with Jacob Appelbaum on Autonomy Cube, a transparent sculpture containing functioning network hardware that routed gallery visitors’ internet traffic through the Tor anonymity network. The sculpture combines Minimalist form, technological infrastructure, privacy tools, and institutional critique. The cube does not merely symbolize privacy; it performs it, transforming the museum space into a node for an alternative communications network.

Orbital Reflector: A Satellite Without Utility

Paglen’s Orbital Reflector was conceived as a nonfunctional satellite intended to become visible from Earth as a slowly moving artificial star. Unlike ordinary satellites, it was not designed for surveillance, communication, military use, or commercial service. Its purpose was entirely aesthetic and speculative, asking: Can space technology be used for something other than utility, profit, or control?

Although technical and political circumstances prevented its full intended realization, the project remains important as an attempt to redirect aerospace infrastructure toward public imagination. The satellite becomes a counter-image: instead of looking down at Earth, it would be looked at; instead of gathering data, it would produce a shared visual event. The prototype invites viewers to imagine machinery in the sky that is not tracking them.

Influence Operations and the Manufacturing of Reality

Paglen’s more recent work has examined propaganda, psychological operations, conspiracy, belief, and the institutional production of perception. His presentations with Pace focused on the lasting cultural effects of military and CIA influence operations, while his 2026 Guggenheim program addressed belief, deception, and conspiratorial structures. (Pace Gallery)

This development expands his inquiry: the hidden system does not merely watch the world, it helps create what people believe the world is. Influence operations use images, narratives, and media channels to make the battlefield entirely perceptual. Power seeks not only to control territory, but to control interpretation. The observer is produced by systems before the encounter begins; the categories through which they interpret a scene have already been trained.

The Horizon Waved, and Nothing Was Certain

Paglen’s 2025 exhibition Cardinals presented photographs of unusual aerial phenomena recorded in the American West over approximately two decades, examining the conditions under which a strange light in the sky triggers public conspiracy and technological awe. (Pace Gallery)

This culminated in his 2026 exhibition The Horizon Waved, and Nothing Was Certain: 2006–2026, which brought together new and earlier landscapes addressing surveillance, secret military infrastructure, machine vision, and unidentified objects in orbit. (Jessica Silverman Gallery) The title captures a central condition of the practice. The horizon traditionally offers orientation, separating land from sky and stabilizing the image. But in Paglen’s work, the horizon shimmers, blurs, or becomes unreliable. The line dividing what is visible from what is hidden begins to move, producing an epistemological instability: What can be known? What can be trusted? What systems are operating outside public awareness?

The Central Contradiction

The deepest contradiction in Paglen’s practice is:

He uses images to expose systems whose power increasingly depends on images that people never see.

The work is documentary and uncertain, beautiful and politically disturbing, evidentiary and incomplete, technically precise and visually blurred, public and concerned with secrecy, humanly viewed and increasingly about machine vision. The work cannot fully escape the systems it investigates; photography relies on optical technology, digital images rely on computation, and research relies on data. That contradiction gives the practice its pressure.

The Viewer’s Position

Paglen positions the viewer as someone attempting to see from a structurally disadvantaged position, given a blur, a streak, a distant point, or a classification grid. The work requires curiosity, contextual knowledge, doubt, and patience. But the viewer also becomes deeply implicated. Many of the systems Paglen studies depend upon ordinary participation: smartphones, social media, cloud services, location data, and online behavior. The observer is also being observed; the user is also a data source. The viewer stands inside the system being examined.

How the Medium Thinks

Paglen’s media are chosen according to the structures under investigation. Photography carries documentary authority while revealing the limits of visibility. Astronomy makes classified objects in orbit publicly perceptible. Geography locates supposedly immaterial power within territory and infrastructure. Machine learning reveals how automated systems classify and act upon images, while sculpture materializes networks at a bodily scale. Writing and research make visible the institutional and technical knowledge that photographs alone cannot communicate. Every medium asks how seeing is organized by power.

What the Practice Has Discovered

Across photography, investigation, engineering, sculpture, and artificial intelligence, Paglen’s work appears to have produced several major discoveries:

1. Invisibility is often manufactured: Systems remain unseen through distance, classification, technical complexity, legal secrecy, and perceptual habit.
2. Landscape is infrastructure: Deserts, oceans, clouds, and skies contain political and technological systems.
3. Visibility is unequally distributed: Power often sees clearly while remaining difficult to see in return.
4. Images increasingly act without human viewers: Machine-readable images can classify, target, predict, and govern.
5. Classification produces reality: Datasets and labels do not merely describe the world; they shape how institutions treat it.
6. Beauty can reveal and conceal power simultaneously: A beautiful photograph can attract attention while risking the aestheticization of surveillance.
7. Digital systems remain physically grounded: The cloud depends upon cables, servers, sensors, satellites, and territory.
8. Counter-images can become counter-infrastructures: Art can sometimes produce alternative uses of technology rather than merely criticizing existing ones.

The Artistic Capacity

Paglen has developed the capacity to make hidden infrastructures perceptible without pretending they have become fully transparent. He can make a blurred landscape reveal secrecy, a satellite trail become a political image, a cloud become a site of machine perception, and a training dataset become cultural history. He shows how a distant drone exposes unequal vision, turns a beach into global communications infrastructure, and transforms a nonfunctional satellite into a speculative public artwork. Research does not merely accompany the photograph; it changes what the photograph is, giving the research perceptual force and the image political specificity.

The World Paglen Has Built

Paglen’s artistic world includes a recurring vocabulary: deserts, horizons, skies, satellites, drones, cables, classified architecture, clouds, datasets, grids, machine vision, surveillance interfaces, strange lights, and hidden networks. These forms belong to one sustained world because they all concern the politics of perception: What can be seen? By whom? For what purpose? With what consequences? The world feels uncanny because ordinary visual experience is surrounded by hidden systems that observe, classify, and act.

Is the Language Still Learning?

Paglen’s visual vocabulary is increasingly recognizable, creating a critical risk where a blurred military landscape, satellite trail, machine-vision grid, or algorithmically processed cloud can now appear immediately familiar. Secrecy and technological complexity can generate a false sense of significance before the specific inquiry has been fully developed. The central question is: Does each new work reveal a new structure of perception and power, or does it repeat the atmosphere of hidden technological menace?

The language remains alive when a new sensing technology changes what visibility means, machine vision alters the role of the image, or influence operations shift the inquiry from observation to belief. His 2025–2026 exhibitions continue expanding from surveillance and AI toward uncertain aerial phenomena, conspiracy, and the production of belief. (Jessica Silverman Gallery) The language weakens when technical opacity becomes a substitute for formal or conceptual discovery; complexity is not automatically depth.

Risks and Limits

• Research overpowering the encounter: The contextual information may be more compelling than the physical image itself.
• Beauty aestheticizing violence: Atmospheric photographs can make military and surveillance systems visually seductive.
• Technical authority becoming default prestige: Specialized knowledge may grant the artist authority that viewers cannot independently evaluate.
• Critique reproducing extraction: The use of datasets and human images may repeat ethical problems involving consent and classification.
• Secrecy becoming romanticized: Hidden systems can acquire mystery and fascination that accidentally overshadow their consequences.
• Machine vision becoming generalized: Different automated systems operate through different goals and risks; “AI” can become too broad a category.
• Exposure without systemic change: Making a system visible does not automatically weaken its real-world function.

These risks define the pressures its strongest projects must confront, ensuring the spatial and technological question remains felt before it is fully explained.

What the Work Expands

Paglen expands what landscape art can address, extending it to orbital space, military secrecy, communications networks, and machine perception. He expands what documentary photography can be, showing not only the subject, but the structural difficulty of seeing it. He shifts image politics from how humans represent one another to how machines classify people and how images trigger automated action, revealing that vision is not simply a natural human faculty, but is increasingly engineered, distributed, automated, privatized, and weaponized.

What Artists Can Learn

The lesson is not to photograph drones, satellites, cables, or computer screens merely because technology feels contemporary. The deeper lesson is to ask what system is shaping experience while remaining absent from ordinary representation. An artist can investigate what the environment hides, what infrastructure enables, what classifications determine, and who possesses the power to observe. Paglen demonstrates how research can become form—distance becomes blur, secrecy becomes partial visibility, classification becomes a grid, and orbital infrastructure becomes a mark across the sky.

His practice also deepens the inquiry into performed identity: a person does not only perform for other people. Increasingly, the person performs for systems that scan, classify, rank, predict, approve, or target, proving that while the machine may completely misunderstand the performance, its misunderstanding can still carry profound real-world consequences.