Trevor Paglen on How Machines Hijacked Visual Culture

A Book That Reframes the Image

Trevor Paglen's new book, collecting a decade of his writings on technology and visual culture, arrives at a moment when the questions it poses have become unavoidable. The artist and geographer, known for his investigative practice exposing the invisible infrastructures of surveillance, argues that two technological revolutions have fundamentally transformed our relationship to images. The rise of computer vision and generative media, he contends, is more consequential than the discovery of perspective or the invention of photography. As The Art Newspaper reports, Paglen's writing is "personal, rigorous and imaginative," asking questions that few in the AI discourse are raising.

The central thesis is that digital images are no longer representational. They have become operational interfaces, servicing corporate or state interests to extract value and exert influence at a scale and granularity that previous image technologies never achieved. The implications extend far beyond the art world. If images are now engineered prompts designed to activate neurological circuits, as Paglen suggests, then our traditional frameworks for understanding visual culture are obsolete.

The book collects essays written over the last decade, allowing readers to trace the evolution of Paglen's thinking as the technologies he describes have advanced. The earliest essays map the terrain of computer vision and its training data, while the later pieces grapple with the implications of generative media that barely existed when the project began. This chronological structure reveals a thinker refining his analysis in real time, adjusting his framework as the technologies he critiques accelerate beyond what anyone predicted.

Computer Vision and the Machine-Readable World

Paglen explains the relevant technology with clarity. Computer vision describes the tools that machines use to translate the visual world into mathematical abstractions, creating machine-readable images that operate invisibly alongside human perception. These systems have achieved mass adoption since the 2010s and now automate vast amounts of the seeing and reading of images, creating an invisible machine-to-machine image world that already dwarfs human visual culture. The technology encompasses facial recognition, license plate reading, and the monitoring of commercial and industrial spaces.

What makes these systems particularly dangerous, in Paglen's analysis, is that they are "extraordinarily intimate instruments of power" that penetrate human lives at scales both smaller and larger than previously possible, while leaving humans mostly out of the analytical loops. The foundational data sets on which these systems are trained encode the biases of their creators. Paglen's interrogation of ImageNet, one of the most influential image training sets, reveals horrifically negligent categorization riddled with racist, gender, and other biases that now infect the digital systems increasingly shaping our world.

The scale of this machine-to-machine image economy is difficult to overstate. Every day, automated systems process billions of images for purposes ranging from targeted advertising to border control to predictive policing. These systems do not see the way humans see. They extract features, classify objects, and assign identities in processes that are opaque to the people they affect. Paglen argues that this invisible image economy has already become the dominant mode of image production and consumption, even though most people remain unaware of its existence.

From the Society of the Spectacle to the Society of Psyop

The book's most speculative essay dives into the history of scientific research into human perception and its manipulation. Paglen traces a path through experiments on kittens, CIA programs, and the techniques of professional magicians, constructing a genealogy of the cognitive technologies now deployed across digital media. In a nod to the theorist Guy Debord, Paglen dubs our current condition the "Society of Psyop," where images function as engineered payloads designed to directly activate target neurological circuits.

This framework represents a significant evolution from earlier critiques of mass media. Where Debord described the spectacle as a system of representation that mediated social relations, Paglen argues that contemporary images operate below the threshold of representation altogether. They are not showing or depicting reality but actively intervening in it. The distinction between an image and a weapon, in this analysis, has become dangerously thin. As contemporary art criticism on platforms like Hyperallergic grapples with the implications of AI-generated images, Paglen's work provides a theoretical foundation that few other voices offer.

The reference to Debord is not merely academic. Debord's Society of the Spectacle (1967) argued that under capitalism, social life had been replaced by its representation, and that authentic experience had been colonized by commodity culture. Paglen's update suggests that this colonization has now penetrated further, operating not at the level of representation but at the level of cognition itself. Where the spectacle mediated experience through images, the psyop bypasses experience entirely, targeting the neurological processes that underpin perception.

Why This Matters Beyond Technology Criticism

Paglen's position as an artist, rather than a technologist or academic, gives the book its distinctive power. His practice involves physically traveling to the sites of surveillance infrastructure, photographing classified installations, and making visible the systems that operate in secret. This embodied, investigative approach grounds his theoretical claims in material reality. The book argues not just that technology has changed images, but that the relationship between images and power has been fundamentally restructured.

The political timing of the book is pointed. Paglen notes that Silicon Valley's right-wing turn contradicts the earlier belief that information technologies were inherently liberating. The same systems that promised democratization now serve what he calls "cognitive injection attacks," cumulative hacks on our collective sense of reality. Against these systems, the only safe spaces may be entirely outside the digital infrastructure. Good luck finding them, Paglen writes. The task of carving out spaces for genuine perception and critical thought falls to us.

The book also connects to broader conversations about AI ethics and accountability that have intensified in recent years. While much of the public debate has focused on issues like bias in algorithmic decision-making and the displacement of human labor, Paglen's critique operates at a more fundamental level. He is not asking how we can make AI systems fairer or more transparent. He is asking whether the entire framework of machine-readable images and automated perception has already transformed the conditions under which fairness and transparency are even possible.

What the Book Means for Art

For the art world, Paglen's thesis raises uncomfortable questions. If images are no longer representational but operational, what happens to the practice of making images? How should museums, galleries, and artists respond to a condition where the dominant image culture is machine-generated and machine-readable? The book does not offer prescriptions, but its analysis suggests that art's role may be precisely to preserve and develop forms of seeing that resist automation.

Paglen's own work, held in collections including SFMOMA and the Whitney Museum, models this resistance. His photographs of classified military installations, his maps of undersea surveillance cables, and his studies of training data sets all insist on the possibility of human perception against the grain of machine vision. The book extends this practice into language, constructing a critical vocabulary for the condition he describes. Whether that vocabulary will be sufficient to counter the forces he names remains an open question, but it is among the most serious attempts yet to articulate what is at stake when machines learn to see.

Paglen's career as an artist has consistently bridged the gap between academic research and public engagement. Trained as a geographer, he has photographed classified military bases from public land, documented the undersea cables that carry global communications, and trained telescopes at the satellites that surveil the planet. His work has been exhibited at the San Francisco Museum of Modern Art, the Metropolitan Museum of Art, and the Smithsonian American Art Museum. This book collects the intellectual framework that underpins that visual practice, offering readers a guided tour of the systems that increasingly mediate what we see and how we see it.

The relevance of Paglen's analysis extends beyond the art world into the foundations of democratic governance. If images are increasingly engineered to bypass rational engagement and target neurological processes directly, then the frameworks that democratic societies rely on for informed decision-making are under threat. The book does not offer policy solutions, but it provides the conceptual tools needed to understand what is happening. For artists, the implications are equally significant. If the dominant image culture is machine-generated and machine-readable, then the practice of making images for human perception becomes an act of resistance, a refusal of the automated image economy that Paglen describes. This is perhaps the book's most important contribution: it reframes artistic practice not as a luxury or a commodity but as a necessary form of resistance against the colonization of human perception.

For readers approaching the book without a background in technology or critical theory, Paglen's prose remains accessible throughout. He explains complex systems in language that avoids both jargon and oversimplification, and his artist's eye for the visual and material dimensions of his subject matter keeps the analysis grounded. The book is not an easy read, but it is a necessary one, and it arrives at a moment when the questions it poses can no longer be deferred. Whether the frameworks it offers will prove adequate to the scale of the transformation it describes is a question that only the coming years will answer. What is certain is that Paglen has identified the central cultural question of our time: what happens to human perception when machines learn to see faster than humans can understand what they are seeing.