<<
SELECTED CLASSES
Course: Artistic Research & Non-Hegemonic AI Training Datasets
(undergraduates and postgraduates)
School of Culture and Society, Leuphana University, 2024-2025.
This course aimed to discuss practices of Artistic Research as instruments for creating non-hegemonic perspectives, understandings, and training data in the fields of arts and technology. How could investigations interested in not fully visible layers of the arts and AI help us not only reduce the problem of opacity, but also create new data to (de)train the historically automated processes inherent in these fields?
At the meetings, researcher and artist Bruno Moreschi discussed a set of artistic practices coordinated by him, but also previous and more historical references, which could help us understand the possible methodologies, gains, and difficulties of using artistic investigation practices in machine learning. The focus of the discussion was always on the data that trained machines, such as images from Large Scale Vision Datasets. Emphasis was placed on the importance of thinking about possible changes not only necessarily in AI itself, but in stages prior to it, such as practices, files, and understandings that had historically not been considered "valid data." In addition to discussing artistic practices for this, we used theoretical materials, almost always from the Global South, especially the Latin American context. In the field of arts and education, the practice of mail art in the contexts of dictatorships in Latin America during the 1960s and 1970s and the process of political literacy through images by Paulo Freire, for example, were explored.
Throughout the semester, students also worked with a printed image viewed from computer vision datasets. They were encouraged to spend time with this material to speculate on new ways to use these training images in ways that did not replicate the tagging and descriptions done (often poorly) by crowd workers on platforms like Amazon Mechanical Turk.
Course: “Images Strike” – Technology, work and art system
(general audience)
Pinacoteca do Estado de São Paulo, Brazil, 2023.
The course explored contemporary technologies from perspectives that highlighted materialities, bodies, and mediations. To achieve this, the art system—especially the museum—served as a space for collections and experiences. The meetings focused on the human labor involved in the training of so-called Artificial Intelligence; the ways of “seeing” from a computational perspective; the limitations, choices, and absences in machine training datasets; and the existence (in the sense of resistance) of non-hegemonic ways of thinking about technology.
At the heart of this course was a commitment to listening. Our primary material came from listening actions carried out with workers at the Pinacoteca do Estado de São Paulo—guards, receptionists, cleaning staff, and others—who shared their experiences and interactions with the museum's technological infrastructure. Their voices were central to our discussions, offering crucial insights into the often-overlooked human dimensions of technology in cultural institutions. As complementary material, we engaged with a selection of artworks from the Pinacoteca collection, as well as texts and videos, to further contextualize and expand our reflections.
Course: demonumenta
(undergraduates and postgraduates, w/ Prof. Giselle Beiguelman).
Faculty of Architecture, Urbanism and Design, University of São Paulo, 2022-2024
The first stage in partnership with Massachusetts Institute of Technology (MIT) MediaLab.
demonumenta is a research and university extension project that proposes a pioneering pedagogical model for developing creative procedures and a critical repertoire on historical heritage. It encompasses historical research, 3D modeling, and AI systems.
Conceived by Prof. Giselle Beiguelman and researcher Bruno Moreschi during the Covid-19 pandemic, this pedagogical project combined online theoretical classes with international researchers (especially from the MIT Media Lab), field research on public monuments in the city of São Paulo (while respecting health protocols), and collaborative online work using technological tools to create 3D models and deconstruct official symbols of the city that often perpetuate colonial narratives.
The project progressed through multiple stages beyond the classes, including the development of an Augmented Reality application for mobile devices, exhibitions, books, academic articles, documentaries, and more.
Course: Art, Technology and the City
(postgraduates, w/ Prof. Giselle Beiguelman).
Faculty of Architecture, Urbanism and Design, University of São Paulo, 2020-2021
In partnership with Professor Giselle Beiguelman, students were encouraged to think critically about the city's technological infrastructure, drawing on both contemporary and historical artistic and activist research as references. In this context, more opaque technological infrastructures, such as AI, were also studied in depth.
Marked by extensive discussions and presentations of students' ongoing projects, the class not only offered a critical perspective on urban infrastructure but also explored new ways of living collectively.
Course: Art History II
(postgraduates, with a focus on photography)
Visiting Professor, Armando Álvares Penteado Foundation (Brazil), 2017
This course offers an overview of the history of art, with a central focus on the construction of visual perception and ways of seeing, particularly in relation to the history of photography. It explores how artistic movements, individual artists, and creative strategies from the early 20th century onward have engaged with the evolving field of images—not only through artistic photography but also through broader practices of image manipulation, reinterpretation, and dissemination.
A key objective of the course is to examine how images are produced, circulated, and transformed across different media and contexts. By analyzing various artistic strategies, students will critically engage with how photography and other image-based practices intersect with larger cultural, social, and technological shifts. Discussions will extend beyond traditional photographic techniques to include experimental and contemporary approaches that challenge conventional understandings of representation and authorship.
The course also emphasizes the need for a more inclusive and diversified approach to art history. Rather than adhering strictly to a periodized or Eurocentric narrative, the syllabus integrates decolonial and feminist perspectives, highlighting artists and movements that have been historically marginalized. Through this lens, students will explore alternative narratives and voices that reshape our understanding of art history, making it more reflective of global and intersectional experiences.
Courses:
Art History II
Art History III
Art History V (undergraduates).
Art Institute, State University of Campinas (Brazil), 2014-2016
During his experience as a guest professor teaching Art History as part of his PhD, Bruno Moreschi not only covered the official course content but also actively engaged with students to develop perspectives and materials in Portuguese that decolonize the field—traditionally centered on a European, white, and male point of view. Throughout these courses, the methodologies applied and the references studied became the primary inspiration for the creation of a free printed material that highlighted the gaps in the main books used to teach Art History. This effort led to the development of the pamphlet History of _rt.
This printed material presents both qualitative and quantitative data on 2,443 artists drawn from 11 books commonly used in undergraduate visual arts courses in Brazil. The aim is to critically assess the narrow perspectives of the official history of art taught today by compiling and comparing fundamental information about the artists featured in these books. An analysis of the data revealed stark imbalances: of the 2,443 artists, only 215 (8.8%) are women, 22 (0.9%) are Black, and 645 (26.3%) are non-European. Among these 645 non-European artists, just 246 are from outside the United States. Additionally, the data show a strong emphasis on painting, with 1,566 painters included in the total count. (See our methodology here.)
The findings were compiled into a pamphlet designed to resemble those distributed at museum entrances, making it accessible and familiar to the public. A total of 13,000 copies were printed in Portuguese and 7,000 in English, and in 2017, the material was freely distributed at museum entrances worldwide. More here.
SELECTED STUDENT SUPERVISIONS
Master Leuphana University, Germany
This Post Has Been Removed: Contente moderation, AI and marginalised communities
Co-supervision.
Content moderation on social media is a widely discussed issue especially given recent debates on X’s new forms of content moderation and Meta’s announcement to pursue a similar route. This thesis aims to explore the topic from the perspective of the users as well as the workers conducting the content moderation, focussing on how it affects marginalised communities. The aim is to analyse power structures at play and how generative AI has complicated the topic.
Master in Arte / Digital Medi, Hochschule für Künste Bremen / University of the Arts Bremen (HfK Bremen), 2024 - ...
Co-supervision. Supervision by Professors Annette Geiger and Andrea Sick.(master).
The growing fusion of the boundaries between the physical and the digital, in addition to the presence of new technologies, tools and languages, drastically alters our experience of form and landscape. In this master's research, researcher Gabriella Gonçallles seeks to understand how digitalization and cultural transformation influence the way we design and experience spaces, and how this changes the meaning we attribute to them.
MIT - Massachusetts Institute of Technology and University of São Paulo, Fulbright US Student Program, 2021-2022
Supervision (master)
DALL-E 2 supports basic input in a number of languages, such as Spanish, German and Portuguese. This is an exciting development for non-English speaking artists and designers who could benefit from using the software. However, prompts given in other languages sometimes lack accuracy and differ in stylistic choices. In this work, Gretchen M. Eggers explored the differences in aesthetic quality and accuracy of results produced by DALL-E 2 from paired text inputs in English and Portuguese.
Undergraduate in Engineering, University of São Paulo, 2020-2022 - PIBIC (Institutional Scientific Initiation Scholarship Program) - CNPQ scholarship.
Supervision (bachelor).
Vinicius Ariel examined the ambivalent dynamics of activism in social media and online platforms. With a group of researchers from areas such as Communication, Visual Arts and Design, Anthropology, Computer Science and Engineering, he analyzed 213,083 images shared on Instagram that are part of the hashtag #MariellePresente, an online political manifestation that arose in response to the assassination of Brazilian councilwoman Marielle Franco in 2018, an unsolved case. After collecting images with a Python programming language script, he used two Computer Vision/Artificial Intelligence tools to read them (Google Cloud Vision and YOLO Darknet). The results showed the capitalist logics inscribed into these technologies and also shed light on the role played by both online activism and data analysis tools. Thus, the consequences of the shift of political movements online became apparent: by helping activism to find its audience, online platforms simultaneously subject its cause to demands of 21st century digital capitalism. His research culminated in this article published in the journal Digital Culture & Society.
Undergraduate in Molecular Science, University of São Paulo, 2020-2022
Supervision (bachelor).
The project audits fake human faces generated by the website This Person Does Not Exist (TPDNE), and discusses how this system can help to perpetuate normativities sustained by a dependency on a limited database. The analyses are centered on the “default generic face”, which we
created by overlapping random samples of fake human faces generated by TPDNE's algorithms. The research was presented at the symposium BeyondFairCV – Beyond Fairness: Towards a Just, Equitable, and Accountable Computer Vision, organized by Timnit Gebru and Emily Denton. It was also published in the proceedings of the same event.
SELECTED CLASSES
(undergraduates and postgraduates)
School of Culture and Society, Leuphana University, 2024-2025.
This course aimed to discuss practices of Artistic Research as instruments for creating non-hegemonic perspectives, understandings, and training data in the fields of arts and technology. How could investigations interested in not fully visible layers of the arts and AI help us not only reduce the problem of opacity, but also create new data to (de)train the historically automated processes inherent in these fields?
At the meetings, researcher and artist Bruno Moreschi discussed a set of artistic practices coordinated by him, but also previous and more historical references, which could help us understand the possible methodologies, gains, and difficulties of using artistic investigation practices in machine learning. The focus of the discussion was always on the data that trained machines, such as images from Large Scale Vision Datasets. Emphasis was placed on the importance of thinking about possible changes not only necessarily in AI itself, but in stages prior to it, such as practices, files, and understandings that had historically not been considered "valid data." In addition to discussing artistic practices for this, we used theoretical materials, almost always from the Global South, especially the Latin American context. In the field of arts and education, the practice of mail art in the contexts of dictatorships in Latin America during the 1960s and 1970s and the process of political literacy through images by Paulo Freire, for example, were explored.
Throughout the semester, students also worked with a printed image viewed from computer vision datasets. They were encouraged to spend time with this material to speculate on new ways to use these training images in ways that did not replicate the tagging and descriptions done (often poorly) by crowd workers on platforms like Amazon Mechanical Turk.
Course: “Images Strike” – Technology, work and art system
(general audience)
Pinacoteca do Estado de São Paulo, Brazil, 2023.
The course explored contemporary technologies from perspectives that highlighted materialities, bodies, and mediations. To achieve this, the art system—especially the museum—served as a space for collections and experiences. The meetings focused on the human labor involved in the training of so-called Artificial Intelligence; the ways of “seeing” from a computational perspective; the limitations, choices, and absences in machine training datasets; and the existence (in the sense of resistance) of non-hegemonic ways of thinking about technology.
At the heart of this course was a commitment to listening. Our primary material came from listening actions carried out with workers at the Pinacoteca do Estado de São Paulo—guards, receptionists, cleaning staff, and others—who shared their experiences and interactions with the museum's technological infrastructure. Their voices were central to our discussions, offering crucial insights into the often-overlooked human dimensions of technology in cultural institutions. As complementary material, we engaged with a selection of artworks from the Pinacoteca collection, as well as texts and videos, to further contextualize and expand our reflections.
Course: demonumenta
(undergraduates and postgraduates, w/ Prof. Giselle Beiguelman).
Faculty of Architecture, Urbanism and Design, University of São Paulo, 2022-2024
The first stage in partnership with Massachusetts Institute of Technology (MIT) MediaLab.
demonumenta is a research and university extension project that proposes a pioneering pedagogical model for developing creative procedures and a critical repertoire on historical heritage. It encompasses historical research, 3D modeling, and AI systems.
Conceived by Prof. Giselle Beiguelman and researcher Bruno Moreschi during the Covid-19 pandemic, this pedagogical project combined online theoretical classes with international researchers (especially from the MIT Media Lab), field research on public monuments in the city of São Paulo (while respecting health protocols), and collaborative online work using technological tools to create 3D models and deconstruct official symbols of the city that often perpetuate colonial narratives.
The project progressed through multiple stages beyond the classes, including the development of an Augmented Reality application for mobile devices, exhibitions, books, academic articles, documentaries, and more.
Course: Art, Technology and the City
(postgraduates, w/ Prof. Giselle Beiguelman).
Faculty of Architecture, Urbanism and Design, University of São Paulo, 2020-2021
In partnership with Professor Giselle Beiguelman, students were encouraged to think critically about the city's technological infrastructure, drawing on both contemporary and historical artistic and activist research as references. In this context, more opaque technological infrastructures, such as AI, were also studied in depth.
Marked by extensive discussions and presentations of students' ongoing projects, the class not only offered a critical perspective on urban infrastructure but also explored new ways of living collectively.
Course: Art History II
(postgraduates, with a focus on photography)
Visiting Professor, Armando Álvares Penteado Foundation (Brazil), 2017
This course offers an overview of the history of art, with a central focus on the construction of visual perception and ways of seeing, particularly in relation to the history of photography. It explores how artistic movements, individual artists, and creative strategies from the early 20th century onward have engaged with the evolving field of images—not only through artistic photography but also through broader practices of image manipulation, reinterpretation, and dissemination.
A key objective of the course is to examine how images are produced, circulated, and transformed across different media and contexts. By analyzing various artistic strategies, students will critically engage with how photography and other image-based practices intersect with larger cultural, social, and technological shifts. Discussions will extend beyond traditional photographic techniques to include experimental and contemporary approaches that challenge conventional understandings of representation and authorship.
The course also emphasizes the need for a more inclusive and diversified approach to art history. Rather than adhering strictly to a periodized or Eurocentric narrative, the syllabus integrates decolonial and feminist perspectives, highlighting artists and movements that have been historically marginalized. Through this lens, students will explore alternative narratives and voices that reshape our understanding of art history, making it more reflective of global and intersectional experiences.
Courses:
Art History II
Art History III
Art History V (undergraduates).
Art Institute, State University of Campinas (Brazil), 2014-2016
During his experience as a guest professor teaching Art History as part of his PhD, Bruno Moreschi not only covered the official course content but also actively engaged with students to develop perspectives and materials in Portuguese that decolonize the field—traditionally centered on a European, white, and male point of view. Throughout these courses, the methodologies applied and the references studied became the primary inspiration for the creation of a free printed material that highlighted the gaps in the main books used to teach Art History. This effort led to the development of the pamphlet History of _rt.
This printed material presents both qualitative and quantitative data on 2,443 artists drawn from 11 books commonly used in undergraduate visual arts courses in Brazil. The aim is to critically assess the narrow perspectives of the official history of art taught today by compiling and comparing fundamental information about the artists featured in these books. An analysis of the data revealed stark imbalances: of the 2,443 artists, only 215 (8.8%) are women, 22 (0.9%) are Black, and 645 (26.3%) are non-European. Among these 645 non-European artists, just 246 are from outside the United States. Additionally, the data show a strong emphasis on painting, with 1,566 painters included in the total count. (See our methodology here.)
The findings were compiled into a pamphlet designed to resemble those distributed at museum entrances, making it accessible and familiar to the public. A total of 13,000 copies were printed in Portuguese and 7,000 in English, and in 2017, the material was freely distributed at museum entrances worldwide. More here.
SELECTED STUDENT SUPERVISIONS
- Anne Lorena Benkowitz
Master Leuphana University, Germany
This Post Has Been Removed: Contente moderation, AI and marginalised communities
Co-supervision.
Content moderation on social media is a widely discussed issue especially given recent debates on X’s new forms of content moderation and Meta’s announcement to pursue a similar route. This thesis aims to explore the topic from the perspective of the users as well as the workers conducting the content moderation, focussing on how it affects marginalised communities. The aim is to analyse power structures at play and how generative AI has complicated the topic.
- Gabriella Gonçallles
Master in Arte / Digital Medi, Hochschule für Künste Bremen / University of the Arts Bremen (HfK Bremen), 2024 - ...
Co-supervision. Supervision by Professors Annette Geiger and Andrea Sick.(master).
The growing fusion of the boundaries between the physical and the digital, in addition to the presence of new technologies, tools and languages, drastically alters our experience of form and landscape. In this master's research, researcher Gabriella Gonçallles seeks to understand how digitalization and cultural transformation influence the way we design and experience spaces, and how this changes the meaning we attribute to them.
- Gretchen Eggers
MIT - Massachusetts Institute of Technology and University of São Paulo, Fulbright US Student Program, 2021-2022
Supervision (master)
DALL-E 2 supports basic input in a number of languages, such as Spanish, German and Portuguese. This is an exciting development for non-English speaking artists and designers who could benefit from using the software. However, prompts given in other languages sometimes lack accuracy and differ in stylistic choices. In this work, Gretchen M. Eggers explored the differences in aesthetic quality and accuracy of results produced by DALL-E 2 from paired text inputs in English and Portuguese.
- Vinicius Ariel Arruda dos Santos
Undergraduate in Engineering, University of São Paulo, 2020-2022 - PIBIC (Institutional Scientific Initiation Scholarship Program) - CNPQ scholarship.
Supervision (bachelor).
Vinicius Ariel examined the ambivalent dynamics of activism in social media and online platforms. With a group of researchers from areas such as Communication, Visual Arts and Design, Anthropology, Computer Science and Engineering, he analyzed 213,083 images shared on Instagram that are part of the hashtag #MariellePresente, an online political manifestation that arose in response to the assassination of Brazilian councilwoman Marielle Franco in 2018, an unsolved case. After collecting images with a Python programming language script, he used two Computer Vision/Artificial Intelligence tools to read them (Google Cloud Vision and YOLO Darknet). The results showed the capitalist logics inscribed into these technologies and also shed light on the role played by both online activism and data analysis tools. Thus, the consequences of the shift of political movements online became apparent: by helping activism to find its audience, online platforms simultaneously subject its cause to demands of 21st century digital capitalism. His research culminated in this article published in the journal Digital Culture & Society.
- Lucas Nunes Sequeira
Undergraduate in Molecular Science, University of São Paulo, 2020-2022
Supervision (bachelor).
The project audits fake human faces generated by the website This Person Does Not Exist (TPDNE), and discusses how this system can help to perpetuate normativities sustained by a dependency on a limited database. The analyses are centered on the “default generic face”, which we
created by overlapping random samples of fake human faces generated by TPDNE's algorithms. The research was presented at the symposium BeyondFairCV – Beyond Fairness: Towards a Just, Equitable, and Accountable Computer Vision, organized by Timnit Gebru and Emily Denton. It was also published in the proceedings of the same event.