ーーーーーーーーーーーーーー
■ 発表者・発表要旨（発表順）
Alexandre Gefen
BIO: Alexandre Gefen, Directeur de Recherche (Full Research Professor) at the CNRS Theory and History of Modern Art and Literature Laboratory (UMR7172, THALIM, CNRS / University Sorbonne Nouvelle – Paris 3- Ecole Normale Supérieure), is a historian of ideas and literature. He is the author of numerous articles and essays on culture, contemporary literature and literary theory. He was one of the pioneers of Digital Humanities in France. He is the director of the Culturia IA research project, which focuses on the history and cultural issues of artificial intelligence. Latest books: Territoires de la non-fiction, Brill, 2020. With Olivier Bessard-Banquy and Sylvie Ducas, Best-sellers. L’industrie du succès, Armand Colin, 2021. L’idée de littérature. De l’art pour l’art aux écritures d’intervention, Corti, 2021. La littérature est une affaire politique, L’Observatoire, 2022. La littérature, une infographie, CNRS éditions, 2022. Créativités artificielles, Les Presses du réel, 2023. Vivre avec ChatGPT, L’Observatoire, 2023.
Session 1
Ai Hisano
“Technology as Practice: Body, Materiality, and Aesthetic Intelligence”
This presentation explores how AI has changed people’s sensory experience, as well as their understanding of surrounding environments through their senses. There has been an increasing interest among engineers in the development of AI technologies that could detect, and reproduce, various sensory stimuli, including smell, taste, and tactility. Such technologies are believed to have the potential to, for example, help businesses create products with new kinds of sensory appeal and enhance people’s sensory experience. But “sensory experience” of whom, and in what context?
As scholars in the history of the senses and related fields have shown, the senses are not merely biological or personal phenomena. Nor are they ahistorical. Sensory experiences are shared experiences embedded in certain cultures. I argue that it is important to look at the senses from social and cultural perspectives and foster what could be called “aesthetic intelligence” to better understand the multifaceted implications of AI technologies in society.
In analyzing the technological impact on the human body, particularly sensations, I propose to look at “technology” not merely as a functional material object but also as practice. Technology does not exist in a vacuum—it has to be situated in a broader historical context. By doing so, we can understand how technology is made and used, by whom, and for what purposes. This “archeology” of technologies helps us see not only their technical components but also their political, social, cultural, and ecological implications.
BIO: Ai Hisano is an associate professor at the Graduate School of Interdisciplinary Information Studies at the University of Tokyo. She specializes in the history of technology, the history of the senses, and business history. She is the author of Visualizing Taste: How Business Changed the Look of What You Eat (Harvard University Press, 2019).
Antonio Somaini
“The Visible and the Sayable. On the Biases of Text-to-Image Models and the Strategies to Counter Them”
The presentation will tackle the question of the biases that appear in the images generated by the three currently most popular text-to-image models (Stable Diffusion, DALL-E 2, and Midjourney) and on the various strategies adopted in order to counter them. Starting from an analysis of the structure, the contents, the sources, and the guiding criteria of the dataset that was used to train Stable Diffusion, LAION-5B, we will show that some of such biases and limits are already inscribed in the training sets. Through concrete examples of images generated with different versions of the three models, we will also show how the companies that manage them have come up with strategies in order to counter such biases. Among them, the choice of eliminating from the training sets images that are considered to be “not safe for work” (NTSF), of adding “hidden prompts” (also called “pre-prompts”) to the users’s prompts, and of introducing different forms of “prompt censorship” (through so-called “banned prompts”). These various strategies show that, in their current stage of development, text-to-image models are a contested field, in which what is visible is strictly dependent on what is sayable. What can be visualized out of the vast spectrum of images that are contained in the latent space of the models as statistical possibilities, depends on what can be written in the words and texts used as prompts. Among the examples that will be tackled in order to analyze these questions, there will be AI-generated images by artists and photographers such as Grégory Chatonsky, Boris Eldagsen, David Fathi, and Hito Steyerl.
BIO: Antonio Somaini is professor of film, media, and visual culture theory at the Université Sorbonne Nouvelle in Paris. He is also a senior member of the Institut Universitaire de France (IUF), working on a research project which tackles the impact of AI technologies on images, photography, cinema, contemporary art and visual culture. Among his latest publications, the article “Algorithmic Images : Artificial Intelligence and Visual Culture” (Grey Room 93, Fall 2023), the book Culture visuelle. Images, regards, médias, dispositifs (with Andrea Pinotti, Les Presses du Réel, 2022), and the editing of the books Repenser le médium. Art contemporain et cinéma (with Larisa Dryansky and Riccardo Venturi, Les Presses du Réel, 2022) and La haute et la basse définition des images (with Francesco Casetti, Mimésis, 2021). In 2020 he has been the chief curator of the exhibition Time Machine: Cinematic Temporalities (Parma, Italie, catalogue published by Skira), and he is currently co-curating the exhibitionLe monde selon l’IA / The World Through AI for the Jeu de Paume museum in Paris.
Galina Shyndriayeva
“Automating Creativity: Artificial Intelligence in Perfumery and Design”
The program for this conference states that “art…enables us to understand the historical depth of the science and technology at the heart of contemporary artificial intelligence.” Creative industries are often at the nexus of art, science, and technology, complicated by a commercial imperative. In this talk, I would like to review some of the salient issues arising from the use of AI in design, first from a general overview of AI in design and secondly from the specific case of AI in the fragrance industry, focusing in particular on automating creativity. As scholars of science and the senses have examined, in twentieth-century perfumery and related industries, there exists an important history of projects to make the subjective objective, that is, of creating systems to classify and measure and predict likes and dislikes, pleasant sensations and unpleasant sensations, and to instrumentatlize individual sensory evaluation. These have implications in how consumers have been imagined, analyzed, and marketed to, and have shaped the sensory experience of everyday life. What sort of analysis and action does the use of AI in fragrance design and evaluation allow? In what ways is it a continuation of previous systems? What does it mean to smell, and to smell good, at this moment with AI? This talk will offer a review of current practices and apprehensions, as well as offer some tentative conclusions from comparison with related historical cases centering on automating creativity.
BIO: Galina Shyndriayeva is currently a Researcher at Musashi University, Tokyo, and a Visiting Researcher of the Interfaculty Initiative in Information Studies at the University of Tokyo. Her current book project, Sensual Science: expertise, craft, and chemical invention in global perfumery, 1900-2000, focuses on the development of the perfume industry in the twentieth century. She holds a PhD in history from King’s College London, MSc from Imperial College London, and a BA from Harvard University, and previously taught at the University of Tokyo as a Project Assistant Professor.
Session 2
Lionel Obadia
“‘Official’ Aesthetics and Narratives of AI in Contrast: Comparing Japan and France”
The collective representations and attitudes towards AI are embedded in a certain symbolism and worldviews that are framed by and framing for the experience of technologies in everyday life (from an anthropological point of view). AI and robotics are therefore entrenched in Political programs and narratives: studying the ways these “official” narratives are exhibited in technological museums (Miraikan in Japan and Cité des Sciences in France) I will discuss the common and contrasting views on AI and technological developments, as a first step towards the deconstruction of the so-called “cultural acceptability” of these technologies in their respective cultural backgrounds.
BIO: Lionel Obadia, PhD in Sociology (1997), was Associate Professor in Ethnology at the University of Lille (1998–2004) and is Full Professor of Anthropology (since 2004) at the University of Lyon, France. He also teaches in other French universities (EHESS, EPHE, SciencePo). He is a specialist of anthropology of religion, Asian religions, and globalization. His works focus on hybridization and cultural/religious transfers. He has conducted fieldworks in France, Europe (on Buddhism in the West), Nepal (on Buddhism and Shamanism), the United States and Israel (on Jewish messianism), and South India (in Auroville). His research now explores the relationships between religions and digital technologies with fieldworks in Europe and Japan. He has published 90 papers in peer-reviewed journals ; 21 books : 12 monographs and 9 edited books ; 17 edition of journal special issues ; 52 book chapters, 107 conferences proceedings, 25 reviews. For the detail of publications see: https://www.researchgate.net/profile/Lionel_Obadia
Yuko Itatsu
“To love and be loved: Tales of the Fictoromantic”
Forty percent of people residing in Japan live in single-person households. One and a half million people in Japan live socially recluse. Japan has entered the “marriage ice age,” where young people are reluctant to get married. One in six young people have romantic feelings towards two-dimensional characters in Japan. More than 4000 men in Japan have “married” two-dimensional characters.
This presentation explores the rhetoric used by those who marry two-dimensional characters and their justification for seeking social acknowledgment of their relationship. By analyzing the tales given by these people, the paper hopes to create a taxonomy of the rhetoric used in such storytelling. As a conceptual reference, the art piece “hybrid couple” by Dutch/Spanish artist Alicia Framis will be cited as it experiments with a human relationship with a hologram. In addition, this paper hopes to explore the correlation between the rise of fictoromantics and the increase in single-person households, especially in technologically advanced countries. What does it say about the societal and cultural backdrop that more people are exploring cross-dimensional relationships with fictional characters? What is it about the pursuit of self-fulfillment that encourages engagement with AI-generated two-dimensional characters? How do fictoromantic relationships fit into the current narratives of couplehood? By addressing these questions, this paper aims to offer insight into the role of AI in the evolving nature of human and computer interaction as it intersects with human desire for relationships.
BIO: Yuko Itatsu is a professor at the Graduate School of Interdisciplinary Information Studies at the University of Tokyo. She is the director of the B’AI Global Forum, which examines the social implications of artificial intelligence, especially as it relates to marginalized populations. She is a social and cultural historian of the 20th century and beyond. She and Ai Hisano co-edited Understanding AI in Society: How technology exerts power (University of Tokyo, 2023; in Japanese).
Pierre Cassou-Noguès & Gwenola Wagon
“Uncanny homes: living in AI”
We’ll be using images we’ve generated on Dall-e and Midjourney to explore the uncanniness that generative AIs sometimes produce. The first project we’ll look at concerns real estate ads. As we know, they all look rather similar. Thus it os not surprising that an algorithm manages to produce new ads, both text and images. However, the images of AI-generated housing take on a particular uncanniness that needs to be questioned. At stake is the exact degree of our familiarity with our dwellings, for which we do not project such definite features as we do for another human, or another living being, but in which we nevertheless project an affective content.
In the second part, we propose to explore more broadly the principle whereby generative AIs defamiliarize us with the familiar through their ability to abstract atmospheres and reproduce them on entirely different contents.
BIO: Pierre Cassou-Noguès is Full Professor in the department of philosophy of University Paris 8. He is co-editor of the journal SubStance, published by John Hopkins Univ. Press. His work concern the role of the imagination in science and technology. His books include Les démons de Gödel (which has been translated in Japanese in 2021), and more recently La bienveillance des machines, The benevolence of machines, Seuil, 2022.
BIO: Gwenola Wagon is an artist and researcher. She is Full Professor at the Sorbonne School of Arts and the University of Paris 1. Through installations, films and books, she imagines alternative and paradoxical narratives for thinking the contemporary digital world. She investigating the space of hyperinformation and Internet infrastructures in collaboration with artist Stéphane Degoutin, with whom she has co-produced numerous pieces, including Cyborgs dans la brume and World Brain, and the book Psychanalyse de l’aéroport international. The links between media development and spiritualist practices are the subject of research projects and exhibitions such as Media Mediums and Haunted by Algorithms with Jeff Guess. After Erewhon and Virusland, two post-cybernetic fables co-directed with the philosopher Pierre Cassou-Noguès, she published the book Planète B , an essay that mixes investigation and fiction in order to apprehend a monster in full expansion.
----------------------------
■ Presenters & Abstracts (presentation order)
Alexandre Gefen
BIO: Alexandre Gefen, Directeur de Recherche (Full Research Professor) at the CNRS Theory and History of Modern Art and Literature Laboratory (UMR7172, THALIM, CNRS / University Sorbonne Nouvelle – Paris 3- Ecole Normale Supérieure), is a historian of ideas and literature. He is the author of numerous articles and essays on culture, contemporary literature and literary theory. He was one of the pioneers of Digital Humanities in France. He is the director of the Culturia IA research project, which focuses on the history and cultural issues of artificial intelligence. Latest books: Territoires de la non-fiction, Brill, 2020. With Olivier Bessard-Banquy and Sylvie Ducas, Best-sellers. L’industrie du succès, Armand Colin, 2021. L’idée de littérature. De l’art pour l’art aux écritures d’intervention, Corti, 2021. La littérature est une affaire politique, L’Observatoire, 2022. La littérature, une infographie, CNRS éditions, 2022. Créativités artificielles, Les Presses du réel, 2023. Vivre avec ChatGPT, L’Observatoire, 2023.
Session 1
Ai Hisano
“Technology as Practice: Body, Materiality, and Aesthetic Intelligence”
This presentation explores how AI has changed people’s sensory experience, as well as their understanding of surrounding environments through their senses. There has been an increasing interest among engineers in the development of AI technologies that could detect, and reproduce, various sensory stimuli, including smell, taste, and tactility. Such technologies are believed to have the potential to, for example, help businesses create products with new kinds of sensory appeal and enhance people’s sensory experience. But “sensory experience” of whom, and in what context?
As scholars in the history of the senses and related fields have shown, the senses are not merely biological or personal phenomena. Nor are they ahistorical. Sensory experiences are shared experiences embedded in certain cultures. I argue that it is important to look at the senses from social and cultural perspectives and foster what could be called “aesthetic intelligence” to better understand the multifaceted implications of AI technologies in society.
In analyzing the technological impact on the human body, particularly sensations, I propose to look at “technology” not merely as a functional material object but also as practice. Technology does not exist in a vacuum—it has to be situated in a broader historical context. By doing so, we can understand how technology is made and used, by whom, and for what purposes. This “archeology” of technologies helps us see not only their technical components but also their political, social, cultural, and ecological implications.
BIO: Ai Hisano is an associate professor at the Graduate School of Interdisciplinary Information Studies at the University of Tokyo. She specializes in the history of technology, the history of the senses, and business history. She is the author of Visualizing Taste: How Business Changed the Look of What You Eat (Harvard University Press, 2019).
Antonio Somaini
“The Visible and the Sayable. On the Biases of Text-to-Image Models and the Strategies to Counter Them”
The presentation will tackle the question of the biases that appear in the images generated by the three currently most popular text-to-image models (Stable Diffusion, DALL-E 2, and Midjourney) and on the various strategies adopted in order to counter them. Starting from an analysis of the structure, the contents, the sources, and the guiding criteria of the dataset that was used to train Stable Diffusion, LAION-5B, we will show that some of such biases and limits are already inscribed in the training sets. Through concrete examples of images generated with different versions of the three models, we will also show how the companies that manage them have come up with strategies in order to counter such biases. Among them, the choice of eliminating from the training sets images that are considered to be “not safe for work” (NTSF), of adding “hidden prompts” (also called “pre-prompts”) to the users’s prompts, and of introducing different forms of “prompt censorship” (through so-called “banned prompts”). These various strategies show that, in their current stage of development, text-to-image models are a contested field, in which what is visible is strictly dependent on what is sayable. What can be visualized out of the vast spectrum of images that are contained in the latent space of the models as statistical possibilities, depends on what can be written in the words and texts used as prompts. Among the examples that will be tackled in order to analyze these questions, there will be AI-generated images by artists and photographers such as Grégory Chatonsky, Boris Eldagsen, David Fathi, and Hito Steyerl.
BIO: Antonio Somaini is professor of film, media, and visual culture theory at the Université Sorbonne Nouvelle in Paris. He is also a senior member of the Institut Universitaire de France (IUF), working on a research project which tackles the impact of AI technologies on images, photography, cinema, contemporary art and visual culture. Among his latest publications, the article “Algorithmic Images : Artificial Intelligence and Visual Culture” (Grey Room 93, Fall 2023), the book Culture visuelle. Images, regards, médias, dispositifs (with Andrea Pinotti, Les Presses du Réel, 2022), and the editing of the books Repenser le médium. Art contemporain et cinéma (with Larisa Dryansky and Riccardo Venturi, Les Presses du Réel, 2022) and La haute et la basse définition des images (with Francesco Casetti, Mimésis, 2021). In 2020 he has been the chief curator of the exhibition Time Machine: Cinematic Temporalities (Parma, Italie, catalogue published by Skira), and he is currently co-curating the exhibitionLe monde selon l’IA / The World Through AI for the Jeu de Paume museum in Paris.
Galina Shyndriayeva
“Automating Creativity: Artificial Intelligence in Perfumery and Design”
The program for this conference states that “art…enables us to understand the historical depth of the science and technology at the heart of contemporary artificial intelligence.” Creative industries are often at the nexus of art, science, and technology, complicated by a commercial imperative. In this talk, I would like to review some of the salient issues arising from the use of AI in design, first from a general overview of AI in design and secondly from the specific case of AI in the fragrance industry, focusing in particular on automating creativity. As scholars of science and the senses have examined, in twentieth-century perfumery and related industries, there exists an important history of projects to make the subjective objective, that is, of creating systems to classify and measure and predict likes and dislikes, pleasant sensations and unpleasant sensations, and to instrumentatlize individual sensory evaluation. These have implications in how consumers have been imagined, analyzed, and marketed to, and have shaped the sensory experience of everyday life. What sort of analysis and action does the use of AI in fragrance design and evaluation allow? In what ways is it a continuation of previous systems? What does it mean to smell, and to smell good, at this moment with AI? This talk will offer a review of current practices and apprehensions, as well as offer some tentative conclusions from comparison with related historical cases centering on automating creativity.
BIO: Galina Shyndriayeva is currently a Researcher at Musashi University, Tokyo, and a Visiting Researcher of the Interfaculty Initiative in Information Studies at the University of Tokyo. Her current book project, Sensual Science: expertise, craft, and chemical invention in global perfumery, 1900-2000, focuses on the development of the perfume industry in the twentieth century. She holds a PhD in history from King’s College London, MSc from Imperial College London, and a BA from Harvard University, and previously taught at the University of Tokyo as a Project Assistant Professor.
Session 2
Lionel Obadia
“‘Official’ Aesthetics and Narratives of AI in Contrast: Comparing Japan and France”
The collective representations and attitudes towards AI are embedded in a certain symbolism and worldviews that are framed by and framing for the experience of technologies in everyday life (from an anthropological point of view). AI and robotics are therefore entrenched in Political programs and narratives: studying the ways these “official” narratives are exhibited in technological museums (Miraikan in Japan and Cité des Sciences in France) I will discuss the common and contrasting views on AI and technological developments, as a first step towards the deconstruction of the so-called “cultural acceptability” of these technologies in their respective cultural backgrounds.
BIO: Lionel Obadia, PhD in Sociology (1997), was Associate Professor in Ethnology at the University of Lille (1998–2004) and is Full Professor of Anthropology (since 2004) at the University of Lyon, France. He also teaches in other French universities (EHESS, EPHE, SciencePo). He is a specialist of anthropology of religion, Asian religions, and globalization. His works focus on hybridization and cultural/religious transfers. He has conducted fieldworks in France, Europe (on Buddhism in the West), Nepal (on Buddhism and Shamanism), the United States and Israel (on Jewish messianism), and South India (in Auroville). His research now explores the relationships between religions and digital technologies with fieldworks in Europe and Japan. He has published 90 papers in peer-reviewed journals ; 21 books : 12 monographs and 9 edited books ; 17 edition of journal special issues ; 52 book chapters, 107 conferences proceedings, 25 reviews. For the detail of publications see: https://www.researchgate.net/profile/Lionel_Obadia
Yuko Itatsu
“To love and be loved: Tales of the Fictoromantic”
Forty percent of people residing in Japan live in single-person households. One and a half million people in Japan live socially recluse. Japan has entered the “marriage ice age,” where young people are reluctant to get married. One in six young people have romantic feelings towards two-dimensional characters in Japan. More than 4000 men in Japan have “married” two-dimensional characters.
This presentation explores the rhetoric used by those who marry two-dimensional characters and their justification for seeking social acknowledgment of their relationship. By analyzing the tales given by these people, the paper hopes to create a taxonomy of the rhetoric used in such storytelling. As a conceptual reference, the art piece “hybrid couple” by Dutch/Spanish artist Alicia Framis will be cited as it experiments with a human relationship with a hologram. In addition, this paper hopes to explore the correlation between the rise of fictoromantics and the increase in single-person households, especially in technologically advanced countries. What does it say about the societal and cultural backdrop that more people are exploring cross-dimensional relationships with fictional characters? What is it about the pursuit of self-fulfillment that encourages engagement with AI-generated two-dimensional characters? How do fictoromantic relationships fit into the current narratives of couplehood? By addressing these questions, this paper aims to offer insight into the role of AI in the evolving nature of human and computer interaction as it intersects with human desire for relationships.
BIO: Yuko Itatsu is a professor at the Graduate School of Interdisciplinary Information Studies at the University of Tokyo. She is the director of the B’AI Global Forum, which examines the social implications of artificial intelligence, especially as it relates to marginalized populations. She is a social and cultural historian of the 20th century and beyond. She and Ai Hisano co-edited Understanding AI in Society: How technology exerts power (University of Tokyo, 2023; in Japanese).
Pierre Cassou-Noguès & Gwenola Wagon
“Uncanny homes: living in AI”
We’ll be using images we’ve generated on Dall-e and Midjourney to explore the uncanniness that generative AIs sometimes produce. The first project we’ll look at concerns real estate ads. As we know, they all look rather similar. Thus it os not surprising that an algorithm manages to produce new ads, both text and images. However, the images of AI-generated housing take on a particular uncanniness that needs to be questioned. At stake is the exact degree of our familiarity with our dwellings, for which we do not project such definite features as we do for another human, or another living being, but in which we nevertheless project an affective content.
In the second part, we propose to explore more broadly the principle whereby generative AIs defamiliarize us with the familiar through their ability to abstract atmospheres and reproduce them on entirely different contents.
BIO: Pierre Cassou-Noguès is Full Professor in the department of philosophy of University Paris 8. He is co-editor of the journal SubStance, published by John Hopkins Univ. Press. His work concern the role of the imagination in science and technology. His books include Les démons de Gödel (which has been translated in Japanese in 2021), and more recently La bienveillance des machines, The benevolence of machines, Seuil, 2022.
BIO: Gwenola Wagon is an artist and researcher. She is Full Professor at the Sorbonne School of Arts and the University of Paris 1. Through installations, films and books, she imagines alternative and paradoxical narratives for thinking the contemporary digital world. She investigating the space of hyperinformation and Internet infrastructures in collaboration with artist Stéphane Degoutin, with whom she has co-produced numerous pieces, including Cyborgs dans la brume and World Brain, and the book Psychanalyse de l’aéroport international. The links between media development and spiritualist practices are the subject of research projects and exhibitions such as Media Mediums and Haunted by Algorithms with Jeff Guess. After Erewhon and Virusland, two post-cybernetic fables co-directed with the philosopher Pierre Cassou-Noguès, she published the book Planète B , an essay that mixes investigation and fiction in order to apprehend a monster in full expansion.
