Découvrez
Collections
Jetons
Échanger
Lancements
Activité
Récompenses
Studio
/
Éléments
Détenteurs
Activité
Agrégation
Réseaux
Conditions d'utilisation
Politique de confidentialité
3 218,69 $
Your browser does not support the video tag.
Evolved Hallucinations (2017 - ongoing)

I've long been frustrated by the literalness with which fields like computer science and computer vision approach the topic of seeing. It's a neat trick to point a camera at a picture and have the caption "hot dog" or "not hot dog" appear on the screen. But if you show Rene Magritte's iconic "The Treachery of Images" to such an object recognition model, the classifier will invariably return the result: "This is a pipe." Something is wrong here.

Visual perception is squishy and slippery, formed by each of our unique biological makeups, our memories, history, culture, and our own subjectivities. Just as a monarch butterfly sees a flower entirely differently than a field mouse, a medieval Spanish farmer sees a comet in an entirely different way than a contemporary architect. An early 20th-century psychoanalyst or semiotician might understand images from a dream quite differently than a present-day cognitive neuroscientist. "Seeing" is a deeply historical, cultural, subjective, and even political affair, profoundly shaped by our sensory and social environs.

With the "Evolved Hallucinations" project, I wanted to see what would happen if we tried to build computer vision models based on a wide range of historical, cultural, and notional worldviews. I began training models on allegorical art, symbolism, and metaphor, using image-vocabularies drawn from literature, philosophy, poetry, folklore, and spiritual traditions. Could I build models that embraced the slipperiness and squishiness of visual perception? What would it mean to build a model designed to "see" the world through the extended allegory of Dante? What might the world look like through the "eyes" of future seaweed on a post-human earth? Or the worldview of a Cassandra-like being, fated to see the future but helpless to change it?

The Evolved Hallucinations is my partial answer to that question.

  • Trevor Paglen
Hallucinations

Unlike Neurons, which delve into the neural network's internal processing stages, this series presents images that the network identifies as representative of fully formed concepts like telescopes, mirrors, and humans.

Each dataset or "corpus" is built around specific taxonomies, such as "The Interpretation of Dreams" which includes multiple image categories within them like “Cocaine”, “False Teeth”, and “A Window”. The variation in image quality, from pixelated abstractions to clear renderings, highlights the challenges of training a GAN and shows how neural networks assign numerical values to pixels based on color intensities.

Instead of adhering to the rigid classifications typical of machine learning, the final generated images evoke the fluid and often contradictory ways humans perceive the world, inviting viewers to reflect on the epistemological impact of AI and its role in shaping our understanding of reality.

Your browser does not support the video tag.
Neurons

A neural network is a bit like a brain with many layers, with each layer adding a new level of understanding. These layers help the network gradually make sense of an image by breaking it down into its subcomponent, or “primitive” parts. The subcomponents of a banana, for example, could consist of two arcs, yellow color gradients, some brown spots, a stem, and so on.

Usually, we ask a GAN to create a picture of something specific, like a cat. But with Neurons, we’re exposing the in-between images of places deep within the hidden layers of the neural network - one of the thousands of abstract images it makes for itself to build a complete understanding of the bigger picture. Thus, we’re exposing the internal processes of a GAN, essentially illustrating what it 'sees' or 'thinks about' while it's learning.

The Neurons were made by combining individual corpus models into “supercorpuses”, and then synthesizing neurons from within those larger models.

Corpuses
The corpuses created for this project represent different ways of thinking about training sets. Every training set and every model contains a worldview. The range of categories in a given training set defines the types of concepts a model trained on it can "see" – and crucially – cannot see. The images in each category give shape to those concepts.
The Aftermath of the First Smart War | 10 classes, 41 images

The first Gulf War is widely considered to be the first war to showcase so many technologies that we now take for granted: global positioning systems, "intelligent" missiles and bombs, advanced "stealth" fighters, remote-control vehicles, and the like. Moreover, it was the first war characterized by the military's intense curation of its media coverage. The combination of then-advanced AI systems and the carefully curated reporting beamed back to the American and international public led media theorists to describe it as a watershed moment, where weapon systems, imaging devices, and military propaganda converged into a single phenomenon. This led French theorist Jean Baudrillard, author of the book "Simulation and Simulacra," to declare: "The Gulf War Did Not Take Place."

The first Gulf War is widely considered to be the first war to showcase so many technologies that we now take for granted: global positioning systems, "intelligent" missiles and bombs, advanced "stealth" fighters, remote-control vehicles, and the like. Moreover, it was the first war characterized by the military's intense curation of its media coverage. The combination of then-advanced AI systems and the carefully curated reporting beamed back to the American and international public led media theorists to describe it as a watershed moment, where weapon systems, imaging devices, and military propaganda converged into a single phenomenon. This led French theorist Jean Baudrillard, author of the book "Simulation and Simulacra," to declare: "The Gulf War Did Not Take Place."

American Predators | 10 classes, 35 images

In a world characterized by ever more intense and ever more invasive forms of surveillance, I think about how so many of the systems we rely on every day (the Internet, city streets, smartphones, etc.) are like technological predators, tracking our every move, our every heartbeat, looking for a moment to take a piece out of us. I wanted to create a model that would see a world through the eyes of prey, a model where literally everything it could see was something that was out to get it.

In a world characterized by ever more intense and ever more invasive forms of surveillance, I think about how so many of the systems we rely on every day (the Internet, city streets, smartphones, etc.) are like technological predators, tracking our every move, our every heartbeat, looking for a moment to take a piece out of us. I wanted to create a model that would see a world through the eyes of prey, a model where literally everything it could see was something that was out to get it.

Cosmic Horror | 10 classes, 60 images

The cosmic horror genre emphasizes a universe that is at once utterly incomprehensible to humans and utterly indifferent to them. To quote the genre's most well-known author:

"The most merciful thing in the world, I think, is the inability of the human mind to correlate all its contents. We live on a placid island of ignorance in the midst of black seas of infinity, and it was not meant that we should voyage far. The sciences, each straining in its own direction, have hitherto harmed us little; but some day the piecing together of dissociated knowledge will open up such terrifying vistas of reality, and of our frightful position therein, that we shall either go mad from the revelation or flee from the deadly light into the peace and safety of a new dark age."

  • H.P. Lovecraft, "The Call of Cthulhu"

The cosmic horror genre emphasizes a universe that is at once utterly incomprehensible to humans and utterly indifferent to them. To quote the genre's most well-known author:

"The most merciful thing in the world, I think, is the inability of the human mind to correlate all its contents. We live on a placid island of ignorance in the midst of black seas of infinity, and it was not meant that we should voyage far. The sciences, each straining in its own direction, have hitherto harmed us little; but some day the piecing together of dissociated knowledge will open up such terrifying vistas of reality, and of our frightful position therein, that we shall either go mad from the revelation or flee from the deadly light into the peace and safety of a new dark age."

  • H.P. Lovecraft, "The Call of Cthulhu"
Evolved Hallucinations

On View at Paris Photo 2024

Trevor Paglen

B.1976, Camp Springs, Maryland

Trevor Paglen is known for investigating the invisible through the visible, with a wide-reaching approach that spans image making, sculpture, investigative journalism, writing, engineering, and numerous other disciplines.

The clandestine and the hidden are revealed in series such as The Black Sites, The Other Night Sky, and Limit Telephotography in which the limits of vision are explored through the histories of landscape photography, abstraction, Romanticism, and technology. Paglen’s investigation into the epistemology of representation can be seen in his Symbology and Code Names series which utilize text, video, object, and image to explore questions surrounding military culture and language. Among his chief concerns are learning how to see the historical moment we live in and developing the means to imagine alternative futures.

Paglen has had numerous one-person exhibitions, including at the Museum of Contemporary Art San Diego (2019); Frankfurter Kunstverein, Frankfurt (2015); Eli & Edythe Broad Art Museum, Michigan State University, East Lansing (2015); Protocinema Istanbul (2013); Van Abbemuseum, Eindhoven, Netherlands (2013); and Vienna Secession (2010). He has participated in group exhibitions at the San Francisco Museum of Modern Art (2008, 2010, 2018); Museo Nacional Centro de Arte Reina Sofía, Madrid (2014); The Metropolitan Museum of Art, New York (2012); Tate Modern, London (2010), and numerous other institutions.

Adversarially Evolved Hallucinations

Trevor Paglen, edited by Anthony Downey

Although often considered to be a fault or a glitch in the system, the event of hallucination is central to the models of image production generated by artificial intelligence (AI). Through mining the latent space of computer vision, Trevor Paglen’s series Adversarially Evolved Hallucinations (2017–ongoing) reveals this phantasmal and hallucinatory domain. In the conversation included in this volume, he discusses how we can think from within these opaque structures and, in turn, questions the frequently inflated claims made on behalf of automated image-production systems. In an accompanying essay, Anthony Downey explores the uncanny realm of algorithmically induced images and proposes that AI, through its generative modelling of the world, invariably estranges us from the present and the future.

Trevor Paglen is a multidisciplinary artist known for blending image-making, sculpture, journalism, and engineering into his work. His art, which explores themes like state secrecy and artificial intelligence, has been exhibited globally, including at the Smithsonian Museum of American Art and the Barbican Centre. An acclaimed author, Paglen’s contributions to investigative journalism and art have been recognized with awards like the Electronic Frontier Foundation’s Pioneer Award and the MacArthur Fellowship.

https://www.sternberg-press.com/product/trevor-paglen/

Trevor Paglen introduces Evolved Hallucinations on X
Part 1: Introduction
See more
Part 2: Seeing
See more
Part 3: Computer Vision
See more
Part 4: Neurons
See more
Part 5: Models
See more
Spheres of Heaven, Terraces of Purgatory, Circles of Hell | 40 classe, 234 images

These models try to imagine a Western medieval worldview, where landscapes, images, objects, and concepts are theological allegories. A vision in which the world of objects is characterized by their theological significance.

These models try to imagine a Western medieval worldview, where landscapes, images, objects, and concepts are theological allegories. A vision in which the world of objects is characterized by their theological significance.

Knight, death, and the devil | 16 classe, 56 images

Albrecht Dürer's Meisterstiche—Knight, Death, and the Devil (1513), Saint Jerome in His Study (1514), and Melencolia I (1514)—are central to the history of allegorical art and were key inspirations for Walter Benjamin's theory of images.

In The Origin of German Tragic Drama (1928), Benjamin proposed that allegory is not merely a symbolic or didactic representation but a fragmented, disjointed depiction that reveals the ruins of meaning rather than a direct, unified interpretation. Allegory operates in a space where things stand in for broader concepts but do not seamlessly embody them.

What does it mean to create a dataset and model that holds the question of meaning and non-meaning in a state akin to a quantum superposition, where both exist and do not exist simultaneously?

Albrecht Dürer's Meisterstiche—Knight, Death, and the Devil (1513), Saint Jerome in His Study (1514), and Melencolia I (1514)—are central to the history of allegorical art and were key inspirations for Walter Benjamin's theory of images.

In The Origin of German Tragic Drama (1928), Benjamin proposed that allegory is not merely a symbolic or didactic representation but a fragmented, disjointed depiction that reveals the ruins of meaning rather than a direct, unified interpretation. Allegory operates in a space where things stand in for broader concepts but do not seamlessly embody them.

What does it mean to create a dataset and model that holds the question of meaning and non-meaning in a state akin to a quantum superposition, where both exist and do not exist simultaneously?

Eye / Machine | 3 classes, 26 images

In the three-part trilogy "Eye/Machine" (2001-2003), the late artist and filmmaker Harun Farocki explored the role of machine vision and automated systems in warfare, surveillance, and everyday life. The films show how machines are trained to "see" through targeting systems, surveillance cameras, and automated production lines—mechanical eyes that reduce the world to calculable, measurable, and actionable data, often without human intervention. He concludes the first film with the statement: 'Thinking of a war of autonomous machines, a war without soldiers like a factory without workers'

In the three-part trilogy "Eye/Machine" (2001-2003), the late artist and filmmaker Harun Farocki explored the role of machine vision and automated systems in warfare, surveillance, and everyday life. The films show how machines are trained to "see" through targeting systems, surveillance cameras, and automated production lines—mechanical eyes that reduce the world to calculable, measurable, and actionable data, often without human intervention. He concludes the first film with the statement: 'Thinking of a war of autonomous machines, a war without soldiers like a factory without workers'

From the Depths | 4 classes, 15 images

Composed of images of deep sea creatures, deep space objects, and the instruments humans have used to explore them, “From the Depths” imagines a world populated by entities by objects and entities that inhabit spaces beyond our everyday perceptions.

Composed of images of deep sea creatures, deep space objects, and the instruments humans have used to explore them, “From the Depths” imagines a world populated by entities by objects and entities that inhabit spaces beyond our everyday perceptions.

The Interpretation of Dreams | 8 classes, 74 images

The "Interpretation of Dreams" draws on Sigmund Freud's exploration of the unconscious mind. For Freud, images in the surreal and fragmented landscape of dream-life stand in for latent desires, repressed memories, and unconscious conflicts manifest. It includes imagery that reflects the dream-work process as understood by Freud —condensation, displacement, and the transformation of unconscious thoughts into visual imagery.

The "Interpretation of Dreams" draws on Sigmund Freud's exploration of the unconscious mind. For Freud, images in the surreal and fragmented landscape of dream-life stand in for latent desires, repressed memories, and unconscious conflicts manifest. It includes imagery that reflects the dream-work process as understood by Freud —condensation, displacement, and the transformation of unconscious thoughts into visual imagery.

Monsters of Capitalism | 6 classes, 66 images

A model that can see – and can only see – monsters that have historically been allegories for different moments in a capitalist cycle. Vampires as capitalists, zombies as consumers or laborers, octopuses as monopolies, mummies as the historical connection between capitalism and colonization, etc.

A model that can see – and can only see – monsters that have historically been allegories for different moments in a capitalist cycle. Vampires as capitalists, zombies as consumers or laborers, octopuses as monopolies, mummies as the historical connection between capitalism and colonization, etc.

Things that Exist Negatively | 6 classes, 57 images

A world of entities or phenomena whose existence is defined by absence, lack, or the space around them, rather than by a positive, tangible presence. These are paradoxical types of things, existing as voids, gaps, or openings that shape reality precisely because they are not "things" in the traditional sense.

A world of entities or phenomena whose existence is defined by absence, lack, or the space around them, rather than by a positive, tangible presence. These are paradoxical types of things, existing as voids, gaps, or openings that shape reality precisely because they are not "things" in the traditional sense.

Omens and Portents | 3 classes, 34 images

A model that sees the world as containing only omens and portents. A model that sees signs of danger, change, or transition. A model that sees in the liminal space between superstition, mythology, and cultural belief.

A model that sees the world as containing only omens and portents. A model that sees signs of danger, change, or transition. A model that sees in the liminal space between superstition, mythology, and cultural belief.

Other Gods | 5 classes, 19 images

"Other Gods" is trained to see and identify deities and divine figures from mythologies around the world. The model operates under the premise that gods, demigods, and divine entities are ever-present, hidden within the world's imagery, and waiting to be revealed. What if the only reality were divine realities?

"Other Gods" is trained to see and identify deities and divine figures from mythologies around the world. The model operates under the premise that gods, demigods, and divine entities are ever-present, hidden within the world's imagery, and waiting to be revealed. What if the only reality were divine realities?

Still Life Objects of the Renaissance | 11 classes, 48 images

A model trained to perceive the world solely through the lens of objects found in Renaissance still life paintings, whose depictions of fruits, flowers, food, vessels, and symbolic items conveyed themes of mortality, wealth, and the transience of life. The model turns the contemporary world into a tableau of Renaissance art.

A model trained to perceive the world solely through the lens of objects found in Renaissance still life paintings, whose depictions of fruits, flowers, food, vessels, and symbolic items conveyed themes of mortality, wealth, and the transience of life. The model turns the contemporary world into a tableau of Renaissance art.

Seaweed Dreams | 5 classes, 57 images

A world in the distant future. The humans are long gone. Organic forms flow through submerged and subconscious landscapes. A model inspired by the anonymously authored collection of poems "A Natural History of Seaweed Dreams."

"Plankton-fed, sleep-drugged eyes cast down in the direction of the sacred." “Manta rays bloom in cosmic night.”

  • Anonymous

A world in the distant future. The humans are long gone. Organic forms flow through submerged and subconscious landscapes. A model inspired by the anonymously authored collection of poems "A Natural History of Seaweed Dreams."

"Plankton-fed, sleep-drugged eyes cast down in the direction of the sacred." “Manta rays bloom in cosmic night.”

  • Anonymous
The Humans | 24 classes, 128 images

This dataset explores the space between human representations and the technologies or extensions we have created. It is eclectic and unpredictable, highlighting both the particularities of human identity and the ways we project ourselves through the images and objects we create. “The Humans” captures a mix of images, ideas, and constructs that serve as reflections or distortions of what it means to be human in a world shaped by technology and fantasy.

This dataset explores the space between human representations and the technologies or extensions we have created. It is eclectic and unpredictable, highlighting both the particularities of human identity and the ways we project ourselves through the images and objects we create. “The Humans” captures a mix of images, ideas, and constructs that serve as reflections or distortions of what it means to be human in a world shaped by technology and fantasy.

Evolved Hallucinations by Trevor Paglen

Par trevorpaglen_minter
Ethereum
1 000
oct. 2024
Art
Par trevorpaglen_minter
Ethereum
1 000
A lancé oct. 2024
Art
Prix plancher
0,06 ETH
1er étage %0 %
Meilleure offre
0,0278 WETH
Volume 24 h0,00 ETH
Volume total3,62 ETH
listée0,7 %
Propriétaires (uniques)128 (12,8 %)

Evolved Hallucinations by Trevor Paglen
Evolved Hallucinations by Trevor Paglen

Par trevorpaglen_minter
Ethereum
1 000
oct. 2024
Art
Par trevorpaglen_minter
Ethereum
1 000
A lancé oct. 2024
Art
Explorer
Éléments
Offres
Détenteurs
Attributs
Activité
À propos

Evolved Hallucinations by Trevor Paglen

Par trevorpaglen_minter
Ethereum
1 000
oct. 2024
Art
Par trevorpaglen_minter
Ethereum
1 000
A lancé oct. 2024
Art
Prix plancher
0,06 ETH
1er étage %0 %
Meilleure offre
0,0278 WETH
Volume 24 h0,00 ETH
Volume total3,62 ETH
listée0,7 %
Propriétaires (uniques)128 (12,8 %)

Evolved Hallucinations by Trevor Paglen
Evolved Hallucinations by Trevor Paglen

Par trevorpaglen_minter
Ethereum
1 000
oct. 2024
Art
Par trevorpaglen_minter
Ethereum
1 000
A lancé oct. 2024
Art
Explorer
Éléments
Offres
Détenteurs
Attributs
Activité
À propos
Your browser does not support the video tag.
Evolved Hallucinations (2017 - ongoing)

I've long been frustrated by the literalness with which fields like computer science and computer vision approach the topic of seeing. It's a neat trick to point a camera at a picture and have the caption "hot dog" or "not hot dog" appear on the screen. But if you show Rene Magritte's iconic "The Treachery of Images" to such an object recognition model, the classifier will invariably return the result: "This is a pipe." Something is wrong here.

Visual perception is squishy and slippery, formed by each of our unique biological makeups, our memories, history, culture, and our own subjectivities. Just as a monarch butterfly sees a flower entirely differently than a field mouse, a medieval Spanish farmer sees a comet in an entirely different way than a contemporary architect. An early 20th-century psychoanalyst or semiotician might understand images from a dream quite differently than a present-day cognitive neuroscientist. "Seeing" is a deeply historical, cultural, subjective, and even political affair, profoundly shaped by our sensory and social environs.

With the "Evolved Hallucinations" project, I wanted to see what would happen if we tried to build computer vision models based on a wide range of historical, cultural, and notional worldviews. I began training models on allegorical art, symbolism, and metaphor, using image-vocabularies drawn from literature, philosophy, poetry, folklore, and spiritual traditions. Could I build models that embraced the slipperiness and squishiness of visual perception? What would it mean to build a model designed to "see" the world through the extended allegory of Dante? What might the world look like through the "eyes" of future seaweed on a post-human earth? Or the worldview of a Cassandra-like being, fated to see the future but helpless to change it?

The Evolved Hallucinations is my partial answer to that question.

  • Trevor Paglen
Hallucinations

Unlike Neurons, which delve into the neural network's internal processing stages, this series presents images that the network identifies as representative of fully formed concepts like telescopes, mirrors, and humans.

Each dataset or "corpus" is built around specific taxonomies, such as "The Interpretation of Dreams" which includes multiple image categories within them like “Cocaine”, “False Teeth”, and “A Window”. The variation in image quality, from pixelated abstractions to clear renderings, highlights the challenges of training a GAN and shows how neural networks assign numerical values to pixels based on color intensities.

Instead of adhering to the rigid classifications typical of machine learning, the final generated images evoke the fluid and often contradictory ways humans perceive the world, inviting viewers to reflect on the epistemological impact of AI and its role in shaping our understanding of reality.

Your browser does not support the video tag.
Neurons

A neural network is a bit like a brain with many layers, with each layer adding a new level of understanding. These layers help the network gradually make sense of an image by breaking it down into its subcomponent, or “primitive” parts. The subcomponents of a banana, for example, could consist of two arcs, yellow color gradients, some brown spots, a stem, and so on.

Usually, we ask a GAN to create a picture of something specific, like a cat. But with Neurons, we’re exposing the in-between images of places deep within the hidden layers of the neural network - one of the thousands of abstract images it makes for itself to build a complete understanding of the bigger picture. Thus, we’re exposing the internal processes of a GAN, essentially illustrating what it 'sees' or 'thinks about' while it's learning.

The Neurons were made by combining individual corpus models into “supercorpuses”, and then synthesizing neurons from within those larger models.

Corpuses
The corpuses created for this project represent different ways of thinking about training sets. Every training set and every model contains a worldview. The range of categories in a given training set defines the types of concepts a model trained on it can "see" – and crucially – cannot see. The images in each category give shape to those concepts.
The Aftermath of the First Smart War | 10 classes, 41 images

The first Gulf War is widely considered to be the first war to showcase so many technologies that we now take for granted: global positioning systems, "intelligent" missiles and bombs, advanced "stealth" fighters, remote-control vehicles, and the like. Moreover, it was the first war characterized by the military's intense curation of its media coverage. The combination of then-advanced AI systems and the carefully curated reporting beamed back to the American and international public led media theorists to describe it as a watershed moment, where weapon systems, imaging devices, and military propaganda converged into a single phenomenon. This led French theorist Jean Baudrillard, author of the book "Simulation and Simulacra," to declare: "The Gulf War Did Not Take Place."

The first Gulf War is widely considered to be the first war to showcase so many technologies that we now take for granted: global positioning systems, "intelligent" missiles and bombs, advanced "stealth" fighters, remote-control vehicles, and the like. Moreover, it was the first war characterized by the military's intense curation of its media coverage. The combination of then-advanced AI systems and the carefully curated reporting beamed back to the American and international public led media theorists to describe it as a watershed moment, where weapon systems, imaging devices, and military propaganda converged into a single phenomenon. This led French theorist Jean Baudrillard, author of the book "Simulation and Simulacra," to declare: "The Gulf War Did Not Take Place."

American Predators | 10 classes, 35 images

In a world characterized by ever more intense and ever more invasive forms of surveillance, I think about how so many of the systems we rely on every day (the Internet, city streets, smartphones, etc.) are like technological predators, tracking our every move, our every heartbeat, looking for a moment to take a piece out of us. I wanted to create a model that would see a world through the eyes of prey, a model where literally everything it could see was something that was out to get it.

In a world characterized by ever more intense and ever more invasive forms of surveillance, I think about how so many of the systems we rely on every day (the Internet, city streets, smartphones, etc.) are like technological predators, tracking our every move, our every heartbeat, looking for a moment to take a piece out of us. I wanted to create a model that would see a world through the eyes of prey, a model where literally everything it could see was something that was out to get it.

Cosmic Horror | 10 classes, 60 images

The cosmic horror genre emphasizes a universe that is at once utterly incomprehensible to humans and utterly indifferent to them. To quote the genre's most well-known author:

"The most merciful thing in the world, I think, is the inability of the human mind to correlate all its contents. We live on a placid island of ignorance in the midst of black seas of infinity, and it was not meant that we should voyage far. The sciences, each straining in its own direction, have hitherto harmed us little; but some day the piecing together of dissociated knowledge will open up such terrifying vistas of reality, and of our frightful position therein, that we shall either go mad from the revelation or flee from the deadly light into the peace and safety of a new dark age."

  • H.P. Lovecraft, "The Call of Cthulhu"

The cosmic horror genre emphasizes a universe that is at once utterly incomprehensible to humans and utterly indifferent to them. To quote the genre's most well-known author:

"The most merciful thing in the world, I think, is the inability of the human mind to correlate all its contents. We live on a placid island of ignorance in the midst of black seas of infinity, and it was not meant that we should voyage far. The sciences, each straining in its own direction, have hitherto harmed us little; but some day the piecing together of dissociated knowledge will open up such terrifying vistas of reality, and of our frightful position therein, that we shall either go mad from the revelation or flee from the deadly light into the peace and safety of a new dark age."

  • H.P. Lovecraft, "The Call of Cthulhu"
Spheres of Heaven, Terraces of Purgatory, Circles of Hell | 40 classe, 234 images

These models try to imagine a Western medieval worldview, where landscapes, images, objects, and concepts are theological allegories. A vision in which the world of objects is characterized by their theological significance.

These models try to imagine a Western medieval worldview, where landscapes, images, objects, and concepts are theological allegories. A vision in which the world of objects is characterized by their theological significance.

Knight, death, and the devil | 16 classe, 56 images

Albrecht Dürer's Meisterstiche—Knight, Death, and the Devil (1513), Saint Jerome in His Study (1514), and Melencolia I (1514)—are central to the history of allegorical art and were key inspirations for Walter Benjamin's theory of images.

In The Origin of German Tragic Drama (1928), Benjamin proposed that allegory is not merely a symbolic or didactic representation but a fragmented, disjointed depiction that reveals the ruins of meaning rather than a direct, unified interpretation. Allegory operates in a space where things stand in for broader concepts but do not seamlessly embody them.

What does it mean to create a dataset and model that holds the question of meaning and non-meaning in a state akin to a quantum superposition, where both exist and do not exist simultaneously?

Albrecht Dürer's Meisterstiche—Knight, Death, and the Devil (1513), Saint Jerome in His Study (1514), and Melencolia I (1514)—are central to the history of allegorical art and were key inspirations for Walter Benjamin's theory of images.

In The Origin of German Tragic Drama (1928), Benjamin proposed that allegory is not merely a symbolic or didactic representation but a fragmented, disjointed depiction that reveals the ruins of meaning rather than a direct, unified interpretation. Allegory operates in a space where things stand in for broader concepts but do not seamlessly embody them.

What does it mean to create a dataset and model that holds the question of meaning and non-meaning in a state akin to a quantum superposition, where both exist and do not exist simultaneously?

Eye / Machine | 3 classes, 26 images

In the three-part trilogy "Eye/Machine" (2001-2003), the late artist and filmmaker Harun Farocki explored the role of machine vision and automated systems in warfare, surveillance, and everyday life. The films show how machines are trained to "see" through targeting systems, surveillance cameras, and automated production lines—mechanical eyes that reduce the world to calculable, measurable, and actionable data, often without human intervention. He concludes the first film with the statement: 'Thinking of a war of autonomous machines, a war without soldiers like a factory without workers'

In the three-part trilogy "Eye/Machine" (2001-2003), the late artist and filmmaker Harun Farocki explored the role of machine vision and automated systems in warfare, surveillance, and everyday life. The films show how machines are trained to "see" through targeting systems, surveillance cameras, and automated production lines—mechanical eyes that reduce the world to calculable, measurable, and actionable data, often without human intervention. He concludes the first film with the statement: 'Thinking of a war of autonomous machines, a war without soldiers like a factory without workers'

From the Depths | 4 classes, 15 images

Composed of images of deep sea creatures, deep space objects, and the instruments humans have used to explore them, “From the Depths” imagines a world populated by entities by objects and entities that inhabit spaces beyond our everyday perceptions.

Composed of images of deep sea creatures, deep space objects, and the instruments humans have used to explore them, “From the Depths” imagines a world populated by entities by objects and entities that inhabit spaces beyond our everyday perceptions.

The Interpretation of Dreams | 8 classes, 74 images

The "Interpretation of Dreams" draws on Sigmund Freud's exploration of the unconscious mind. For Freud, images in the surreal and fragmented landscape of dream-life stand in for latent desires, repressed memories, and unconscious conflicts manifest. It includes imagery that reflects the dream-work process as understood by Freud —condensation, displacement, and the transformation of unconscious thoughts into visual imagery.

The "Interpretation of Dreams" draws on Sigmund Freud's exploration of the unconscious mind. For Freud, images in the surreal and fragmented landscape of dream-life stand in for latent desires, repressed memories, and unconscious conflicts manifest. It includes imagery that reflects the dream-work process as understood by Freud —condensation, displacement, and the transformation of unconscious thoughts into visual imagery.

Monsters of Capitalism | 6 classes, 66 images

A model that can see – and can only see – monsters that have historically been allegories for different moments in a capitalist cycle. Vampires as capitalists, zombies as consumers or laborers, octopuses as monopolies, mummies as the historical connection between capitalism and colonization, etc.

A model that can see – and can only see – monsters that have historically been allegories for different moments in a capitalist cycle. Vampires as capitalists, zombies as consumers or laborers, octopuses as monopolies, mummies as the historical connection between capitalism and colonization, etc.

Things that Exist Negatively | 6 classes, 57 images

A world of entities or phenomena whose existence is defined by absence, lack, or the space around them, rather than by a positive, tangible presence. These are paradoxical types of things, existing as voids, gaps, or openings that shape reality precisely because they are not "things" in the traditional sense.

A world of entities or phenomena whose existence is defined by absence, lack, or the space around them, rather than by a positive, tangible presence. These are paradoxical types of things, existing as voids, gaps, or openings that shape reality precisely because they are not "things" in the traditional sense.

Omens and Portents | 3 classes, 34 images

A model that sees the world as containing only omens and portents. A model that sees signs of danger, change, or transition. A model that sees in the liminal space between superstition, mythology, and cultural belief.

A model that sees the world as containing only omens and portents. A model that sees signs of danger, change, or transition. A model that sees in the liminal space between superstition, mythology, and cultural belief.

Other Gods | 5 classes, 19 images

"Other Gods" is trained to see and identify deities and divine figures from mythologies around the world. The model operates under the premise that gods, demigods, and divine entities are ever-present, hidden within the world's imagery, and waiting to be revealed. What if the only reality were divine realities?

"Other Gods" is trained to see and identify deities and divine figures from mythologies around the world. The model operates under the premise that gods, demigods, and divine entities are ever-present, hidden within the world's imagery, and waiting to be revealed. What if the only reality were divine realities?

Still Life Objects of the Renaissance | 11 classes, 48 images

A model trained to perceive the world solely through the lens of objects found in Renaissance still life paintings, whose depictions of fruits, flowers, food, vessels, and symbolic items conveyed themes of mortality, wealth, and the transience of life. The model turns the contemporary world into a tableau of Renaissance art.

A model trained to perceive the world solely through the lens of objects found in Renaissance still life paintings, whose depictions of fruits, flowers, food, vessels, and symbolic items conveyed themes of mortality, wealth, and the transience of life. The model turns the contemporary world into a tableau of Renaissance art.

Seaweed Dreams | 5 classes, 57 images

A world in the distant future. The humans are long gone. Organic forms flow through submerged and subconscious landscapes. A model inspired by the anonymously authored collection of poems "A Natural History of Seaweed Dreams."

"Plankton-fed, sleep-drugged eyes cast down in the direction of the sacred." “Manta rays bloom in cosmic night.”

  • Anonymous

A world in the distant future. The humans are long gone. Organic forms flow through submerged and subconscious landscapes. A model inspired by the anonymously authored collection of poems "A Natural History of Seaweed Dreams."

"Plankton-fed, sleep-drugged eyes cast down in the direction of the sacred." “Manta rays bloom in cosmic night.”

  • Anonymous
The Humans | 24 classes, 128 images

This dataset explores the space between human representations and the technologies or extensions we have created. It is eclectic and unpredictable, highlighting both the particularities of human identity and the ways we project ourselves through the images and objects we create. “The Humans” captures a mix of images, ideas, and constructs that serve as reflections or distortions of what it means to be human in a world shaped by technology and fantasy.

This dataset explores the space between human representations and the technologies or extensions we have created. It is eclectic and unpredictable, highlighting both the particularities of human identity and the ways we project ourselves through the images and objects we create. “The Humans” captures a mix of images, ideas, and constructs that serve as reflections or distortions of what it means to be human in a world shaped by technology and fantasy.

Evolved Hallucinations

On View at Paris Photo 2024

Trevor Paglen

B.1976, Camp Springs, Maryland

Trevor Paglen is known for investigating the invisible through the visible, with a wide-reaching approach that spans image making, sculpture, investigative journalism, writing, engineering, and numerous other disciplines.

The clandestine and the hidden are revealed in series such as The Black Sites, The Other Night Sky, and Limit Telephotography in which the limits of vision are explored through the histories of landscape photography, abstraction, Romanticism, and technology. Paglen’s investigation into the epistemology of representation can be seen in his Symbology and Code Names series which utilize text, video, object, and image to explore questions surrounding military culture and language. Among his chief concerns are learning how to see the historical moment we live in and developing the means to imagine alternative futures.

Paglen has had numerous one-person exhibitions, including at the Museum of Contemporary Art San Diego (2019); Frankfurter Kunstverein, Frankfurt (2015); Eli & Edythe Broad Art Museum, Michigan State University, East Lansing (2015); Protocinema Istanbul (2013); Van Abbemuseum, Eindhoven, Netherlands (2013); and Vienna Secession (2010). He has participated in group exhibitions at the San Francisco Museum of Modern Art (2008, 2010, 2018); Museo Nacional Centro de Arte Reina Sofía, Madrid (2014); The Metropolitan Museum of Art, New York (2012); Tate Modern, London (2010), and numerous other institutions.

Adversarially Evolved Hallucinations

Trevor Paglen, edited by Anthony Downey

Although often considered to be a fault or a glitch in the system, the event of hallucination is central to the models of image production generated by artificial intelligence (AI). Through mining the latent space of computer vision, Trevor Paglen’s series Adversarially Evolved Hallucinations (2017–ongoing) reveals this phantasmal and hallucinatory domain. In the conversation included in this volume, he discusses how we can think from within these opaque structures and, in turn, questions the frequently inflated claims made on behalf of automated image-production systems. In an accompanying essay, Anthony Downey explores the uncanny realm of algorithmically induced images and proposes that AI, through its generative modelling of the world, invariably estranges us from the present and the future.

Trevor Paglen is a multidisciplinary artist known for blending image-making, sculpture, journalism, and engineering into his work. His art, which explores themes like state secrecy and artificial intelligence, has been exhibited globally, including at the Smithsonian Museum of American Art and the Barbican Centre. An acclaimed author, Paglen’s contributions to investigative journalism and art have been recognized with awards like the Electronic Frontier Foundation’s Pioneer Award and the MacArthur Fellowship.

https://www.sternberg-press.com/product/trevor-paglen/

Trevor Paglen introduces Evolved Hallucinations on X
Part 1: Introduction
See more
Part 2: Seeing
See more
Part 3: Computer Vision
See more
Part 4: Neurons
See more
Part 5: Models
See more
Your browser does not support the video tag.
Evolved Hallucinations (2017 - ongoing)

I've long been frustrated by the literalness with which fields like computer science and computer vision approach the topic of seeing. It's a neat trick to point a camera at a picture and have the caption "hot dog" or "not hot dog" appear on the screen. But if you show Rene Magritte's iconic "The Treachery of Images" to such an object recognition model, the classifier will invariably return the result: "This is a pipe." Something is wrong here.

Visual perception is squishy and slippery, formed by each of our unique biological makeups, our memories, history, culture, and our own subjectivities. Just as a monarch butterfly sees a flower entirely differently than a field mouse, a medieval Spanish farmer sees a comet in an entirely different way than a contemporary architect. An early 20th-century psychoanalyst or semiotician might understand images from a dream quite differently than a present-day cognitive neuroscientist. "Seeing" is a deeply historical, cultural, subjective, and even political affair, profoundly shaped by our sensory and social environs.

With the "Evolved Hallucinations" project, I wanted to see what would happen if we tried to build computer vision models based on a wide range of historical, cultural, and notional worldviews. I began training models on allegorical art, symbolism, and metaphor, using image-vocabularies drawn from literature, philosophy, poetry, folklore, and spiritual traditions. Could I build models that embraced the slipperiness and squishiness of visual perception? What would it mean to build a model designed to "see" the world through the extended allegory of Dante? What might the world look like through the "eyes" of future seaweed on a post-human earth? Or the worldview of a Cassandra-like being, fated to see the future but helpless to change it?

The Evolved Hallucinations is my partial answer to that question.

  • Trevor Paglen
Hallucinations

Unlike Neurons, which delve into the neural network's internal processing stages, this series presents images that the network identifies as representative of fully formed concepts like telescopes, mirrors, and humans.

Each dataset or "corpus" is built around specific taxonomies, such as "The Interpretation of Dreams" which includes multiple image categories within them like “Cocaine”, “False Teeth”, and “A Window”. The variation in image quality, from pixelated abstractions to clear renderings, highlights the challenges of training a GAN and shows how neural networks assign numerical values to pixels based on color intensities.

Instead of adhering to the rigid classifications typical of machine learning, the final generated images evoke the fluid and often contradictory ways humans perceive the world, inviting viewers to reflect on the epistemological impact of AI and its role in shaping our understanding of reality.

Your browser does not support the video tag.
Neurons

A neural network is a bit like a brain with many layers, with each layer adding a new level of understanding. These layers help the network gradually make sense of an image by breaking it down into its subcomponent, or “primitive” parts. The subcomponents of a banana, for example, could consist of two arcs, yellow color gradients, some brown spots, a stem, and so on.

Usually, we ask a GAN to create a picture of something specific, like a cat. But with Neurons, we’re exposing the in-between images of places deep within the hidden layers of the neural network - one of the thousands of abstract images it makes for itself to build a complete understanding of the bigger picture. Thus, we’re exposing the internal processes of a GAN, essentially illustrating what it 'sees' or 'thinks about' while it's learning.

The Neurons were made by combining individual corpus models into “supercorpuses”, and then synthesizing neurons from within those larger models.

Corpuses
The corpuses created for this project represent different ways of thinking about training sets. Every training set and every model contains a worldview. The range of categories in a given training set defines the types of concepts a model trained on it can "see" – and crucially – cannot see. The images in each category give shape to those concepts.
The Aftermath of the First Smart War | 10 classes, 41 images

The first Gulf War is widely considered to be the first war to showcase so many technologies that we now take for granted: global positioning systems, "intelligent" missiles and bombs, advanced "stealth" fighters, remote-control vehicles, and the like. Moreover, it was the first war characterized by the military's intense curation of its media coverage. The combination of then-advanced AI systems and the carefully curated reporting beamed back to the American and international public led media theorists to describe it as a watershed moment, where weapon systems, imaging devices, and military propaganda converged into a single phenomenon. This led French theorist Jean Baudrillard, author of the book "Simulation and Simulacra," to declare: "The Gulf War Did Not Take Place."

The first Gulf War is widely considered to be the first war to showcase so many technologies that we now take for granted: global positioning systems, "intelligent" missiles and bombs, advanced "stealth" fighters, remote-control vehicles, and the like. Moreover, it was the first war characterized by the military's intense curation of its media coverage. The combination of then-advanced AI systems and the carefully curated reporting beamed back to the American and international public led media theorists to describe it as a watershed moment, where weapon systems, imaging devices, and military propaganda converged into a single phenomenon. This led French theorist Jean Baudrillard, author of the book "Simulation and Simulacra," to declare: "The Gulf War Did Not Take Place."

American Predators | 10 classes, 35 images

In a world characterized by ever more intense and ever more invasive forms of surveillance, I think about how so many of the systems we rely on every day (the Internet, city streets, smartphones, etc.) are like technological predators, tracking our every move, our every heartbeat, looking for a moment to take a piece out of us. I wanted to create a model that would see a world through the eyes of prey, a model where literally everything it could see was something that was out to get it.

In a world characterized by ever more intense and ever more invasive forms of surveillance, I think about how so many of the systems we rely on every day (the Internet, city streets, smartphones, etc.) are like technological predators, tracking our every move, our every heartbeat, looking for a moment to take a piece out of us. I wanted to create a model that would see a world through the eyes of prey, a model where literally everything it could see was something that was out to get it.

Cosmic Horror | 10 classes, 60 images

The cosmic horror genre emphasizes a universe that is at once utterly incomprehensible to humans and utterly indifferent to them. To quote the genre's most well-known author:

"The most merciful thing in the world, I think, is the inability of the human mind to correlate all its contents. We live on a placid island of ignorance in the midst of black seas of infinity, and it was not meant that we should voyage far. The sciences, each straining in its own direction, have hitherto harmed us little; but some day the piecing together of dissociated knowledge will open up such terrifying vistas of reality, and of our frightful position therein, that we shall either go mad from the revelation or flee from the deadly light into the peace and safety of a new dark age."

  • H.P. Lovecraft, "The Call of Cthulhu"

The cosmic horror genre emphasizes a universe that is at once utterly incomprehensible to humans and utterly indifferent to them. To quote the genre's most well-known author:

"The most merciful thing in the world, I think, is the inability of the human mind to correlate all its contents. We live on a placid island of ignorance in the midst of black seas of infinity, and it was not meant that we should voyage far. The sciences, each straining in its own direction, have hitherto harmed us little; but some day the piecing together of dissociated knowledge will open up such terrifying vistas of reality, and of our frightful position therein, that we shall either go mad from the revelation or flee from the deadly light into the peace and safety of a new dark age."

  • H.P. Lovecraft, "The Call of Cthulhu"
Spheres of Heaven, Terraces of Purgatory, Circles of Hell | 40 classe, 234 images

These models try to imagine a Western medieval worldview, where landscapes, images, objects, and concepts are theological allegories. A vision in which the world of objects is characterized by their theological significance.

These models try to imagine a Western medieval worldview, where landscapes, images, objects, and concepts are theological allegories. A vision in which the world of objects is characterized by their theological significance.

Knight, death, and the devil | 16 classe, 56 images

Albrecht Dürer's Meisterstiche—Knight, Death, and the Devil (1513), Saint Jerome in His Study (1514), and Melencolia I (1514)—are central to the history of allegorical art and were key inspirations for Walter Benjamin's theory of images.

In The Origin of German Tragic Drama (1928), Benjamin proposed that allegory is not merely a symbolic or didactic representation but a fragmented, disjointed depiction that reveals the ruins of meaning rather than a direct, unified interpretation. Allegory operates in a space where things stand in for broader concepts but do not seamlessly embody them.

What does it mean to create a dataset and model that holds the question of meaning and non-meaning in a state akin to a quantum superposition, where both exist and do not exist simultaneously?

Albrecht Dürer's Meisterstiche—Knight, Death, and the Devil (1513), Saint Jerome in His Study (1514), and Melencolia I (1514)—are central to the history of allegorical art and were key inspirations for Walter Benjamin's theory of images.

In The Origin of German Tragic Drama (1928), Benjamin proposed that allegory is not merely a symbolic or didactic representation but a fragmented, disjointed depiction that reveals the ruins of meaning rather than a direct, unified interpretation. Allegory operates in a space where things stand in for broader concepts but do not seamlessly embody them.

What does it mean to create a dataset and model that holds the question of meaning and non-meaning in a state akin to a quantum superposition, where both exist and do not exist simultaneously?

Eye / Machine | 3 classes, 26 images

In the three-part trilogy "Eye/Machine" (2001-2003), the late artist and filmmaker Harun Farocki explored the role of machine vision and automated systems in warfare, surveillance, and everyday life. The films show how machines are trained to "see" through targeting systems, surveillance cameras, and automated production lines—mechanical eyes that reduce the world to calculable, measurable, and actionable data, often without human intervention. He concludes the first film with the statement: 'Thinking of a war of autonomous machines, a war without soldiers like a factory without workers'

In the three-part trilogy "Eye/Machine" (2001-2003), the late artist and filmmaker Harun Farocki explored the role of machine vision and automated systems in warfare, surveillance, and everyday life. The films show how machines are trained to "see" through targeting systems, surveillance cameras, and automated production lines—mechanical eyes that reduce the world to calculable, measurable, and actionable data, often without human intervention. He concludes the first film with the statement: 'Thinking of a war of autonomous machines, a war without soldiers like a factory without workers'

From the Depths | 4 classes, 15 images

Composed of images of deep sea creatures, deep space objects, and the instruments humans have used to explore them, “From the Depths” imagines a world populated by entities by objects and entities that inhabit spaces beyond our everyday perceptions.

Composed of images of deep sea creatures, deep space objects, and the instruments humans have used to explore them, “From the Depths” imagines a world populated by entities by objects and entities that inhabit spaces beyond our everyday perceptions.

The Interpretation of Dreams | 8 classes, 74 images

The "Interpretation of Dreams" draws on Sigmund Freud's exploration of the unconscious mind. For Freud, images in the surreal and fragmented landscape of dream-life stand in for latent desires, repressed memories, and unconscious conflicts manifest. It includes imagery that reflects the dream-work process as understood by Freud —condensation, displacement, and the transformation of unconscious thoughts into visual imagery.

The "Interpretation of Dreams" draws on Sigmund Freud's exploration of the unconscious mind. For Freud, images in the surreal and fragmented landscape of dream-life stand in for latent desires, repressed memories, and unconscious conflicts manifest. It includes imagery that reflects the dream-work process as understood by Freud —condensation, displacement, and the transformation of unconscious thoughts into visual imagery.

Monsters of Capitalism | 6 classes, 66 images

A model that can see – and can only see – monsters that have historically been allegories for different moments in a capitalist cycle. Vampires as capitalists, zombies as consumers or laborers, octopuses as monopolies, mummies as the historical connection between capitalism and colonization, etc.

A model that can see – and can only see – monsters that have historically been allegories for different moments in a capitalist cycle. Vampires as capitalists, zombies as consumers or laborers, octopuses as monopolies, mummies as the historical connection between capitalism and colonization, etc.

Things that Exist Negatively | 6 classes, 57 images

A world of entities or phenomena whose existence is defined by absence, lack, or the space around them, rather than by a positive, tangible presence. These are paradoxical types of things, existing as voids, gaps, or openings that shape reality precisely because they are not "things" in the traditional sense.

A world of entities or phenomena whose existence is defined by absence, lack, or the space around them, rather than by a positive, tangible presence. These are paradoxical types of things, existing as voids, gaps, or openings that shape reality precisely because they are not "things" in the traditional sense.

Omens and Portents | 3 classes, 34 images

A model that sees the world as containing only omens and portents. A model that sees signs of danger, change, or transition. A model that sees in the liminal space between superstition, mythology, and cultural belief.

A model that sees the world as containing only omens and portents. A model that sees signs of danger, change, or transition. A model that sees in the liminal space between superstition, mythology, and cultural belief.

Other Gods | 5 classes, 19 images

"Other Gods" is trained to see and identify deities and divine figures from mythologies around the world. The model operates under the premise that gods, demigods, and divine entities are ever-present, hidden within the world's imagery, and waiting to be revealed. What if the only reality were divine realities?

"Other Gods" is trained to see and identify deities and divine figures from mythologies around the world. The model operates under the premise that gods, demigods, and divine entities are ever-present, hidden within the world's imagery, and waiting to be revealed. What if the only reality were divine realities?

Still Life Objects of the Renaissance | 11 classes, 48 images

A model trained to perceive the world solely through the lens of objects found in Renaissance still life paintings, whose depictions of fruits, flowers, food, vessels, and symbolic items conveyed themes of mortality, wealth, and the transience of life. The model turns the contemporary world into a tableau of Renaissance art.

A model trained to perceive the world solely through the lens of objects found in Renaissance still life paintings, whose depictions of fruits, flowers, food, vessels, and symbolic items conveyed themes of mortality, wealth, and the transience of life. The model turns the contemporary world into a tableau of Renaissance art.

Seaweed Dreams | 5 classes, 57 images

A world in the distant future. The humans are long gone. Organic forms flow through submerged and subconscious landscapes. A model inspired by the anonymously authored collection of poems "A Natural History of Seaweed Dreams."

"Plankton-fed, sleep-drugged eyes cast down in the direction of the sacred." “Manta rays bloom in cosmic night.”

  • Anonymous

A world in the distant future. The humans are long gone. Organic forms flow through submerged and subconscious landscapes. A model inspired by the anonymously authored collection of poems "A Natural History of Seaweed Dreams."

"Plankton-fed, sleep-drugged eyes cast down in the direction of the sacred." “Manta rays bloom in cosmic night.”

  • Anonymous
The Humans | 24 classes, 128 images

This dataset explores the space between human representations and the technologies or extensions we have created. It is eclectic and unpredictable, highlighting both the particularities of human identity and the ways we project ourselves through the images and objects we create. “The Humans” captures a mix of images, ideas, and constructs that serve as reflections or distortions of what it means to be human in a world shaped by technology and fantasy.

This dataset explores the space between human representations and the technologies or extensions we have created. It is eclectic and unpredictable, highlighting both the particularities of human identity and the ways we project ourselves through the images and objects we create. “The Humans” captures a mix of images, ideas, and constructs that serve as reflections or distortions of what it means to be human in a world shaped by technology and fantasy.

Evolved Hallucinations

On View at Paris Photo 2024

Trevor Paglen

B.1976, Camp Springs, Maryland

Trevor Paglen is known for investigating the invisible through the visible, with a wide-reaching approach that spans image making, sculpture, investigative journalism, writing, engineering, and numerous other disciplines.

The clandestine and the hidden are revealed in series such as The Black Sites, The Other Night Sky, and Limit Telephotography in which the limits of vision are explored through the histories of landscape photography, abstraction, Romanticism, and technology. Paglen’s investigation into the epistemology of representation can be seen in his Symbology and Code Names series which utilize text, video, object, and image to explore questions surrounding military culture and language. Among his chief concerns are learning how to see the historical moment we live in and developing the means to imagine alternative futures.

Paglen has had numerous one-person exhibitions, including at the Museum of Contemporary Art San Diego (2019); Frankfurter Kunstverein, Frankfurt (2015); Eli & Edythe Broad Art Museum, Michigan State University, East Lansing (2015); Protocinema Istanbul (2013); Van Abbemuseum, Eindhoven, Netherlands (2013); and Vienna Secession (2010). He has participated in group exhibitions at the San Francisco Museum of Modern Art (2008, 2010, 2018); Museo Nacional Centro de Arte Reina Sofía, Madrid (2014); The Metropolitan Museum of Art, New York (2012); Tate Modern, London (2010), and numerous other institutions.

Adversarially Evolved Hallucinations

Trevor Paglen, edited by Anthony Downey

Although often considered to be a fault or a glitch in the system, the event of hallucination is central to the models of image production generated by artificial intelligence (AI). Through mining the latent space of computer vision, Trevor Paglen’s series Adversarially Evolved Hallucinations (2017–ongoing) reveals this phantasmal and hallucinatory domain. In the conversation included in this volume, he discusses how we can think from within these opaque structures and, in turn, questions the frequently inflated claims made on behalf of automated image-production systems. In an accompanying essay, Anthony Downey explores the uncanny realm of algorithmically induced images and proposes that AI, through its generative modelling of the world, invariably estranges us from the present and the future.

Trevor Paglen is a multidisciplinary artist known for blending image-making, sculpture, journalism, and engineering into his work. His art, which explores themes like state secrecy and artificial intelligence, has been exhibited globally, including at the Smithsonian Museum of American Art and the Barbican Centre. An acclaimed author, Paglen’s contributions to investigative journalism and art have been recognized with awards like the Electronic Frontier Foundation’s Pioneer Award and the MacArthur Fellowship.

https://www.sternberg-press.com/product/trevor-paglen/

Trevor Paglen introduces Evolved Hallucinations on X
Part 1: Introduction
See more
Part 2: Seeing
See more
Part 3: Computer Vision
See more
Part 4: Neurons
See more
Part 5: Models
See more