The Workshop on Cultures of AI and AI for Culture to be held virtually as part of NeurIPS 2022 on December 9. This is a cross-disciplinary workshop with experts in AI/ML and social sciences discussing how AI influences culture and how culture influences AI.
Description
AI development and deployment is shaped by the particular perspectives of the people and institutions it emerges from. Research in AI and ML have shown how human values and biases are baked into AI for instance through annotation practices, training data pipelines, model design choices and through larger organizational incentives. At the same time, AI-driven technologies are also actively shaping human experiences through art, music and language production tools which curate content and shape consumption patterns.
This relationship between AI and such human processes suggests a relationship between AI and what we can broadly call ‘culture’: a set of worldviews, beliefs and social practices. Thus, knowing how AI influences culture (e.g. through content creation or recommendation tools) and how culture influences AI (e.g. through cultural values implicit in systems) becomes an important topic for AI research and development. While nascent conversations in the field have started to question the impact of AI-driven tools on global cultures and the western-centric cultural biases embedded within AI development, foundational research in AI/ML lacks a deeper understanding of what culture is and the ways in which it relates to AI research and development.
Producing culturally sensitive, responsive, and competent AI systems requires us to think about the cultures shaping AI production and about the impact these systems have on global cultures. Without understanding the role of the cultures shaping AI-driven technology, we can not fully respond to or diagnose the biases we are unconsciously globalizing through AI development. Complexifying our understanding of culture is also increasingly salient as AI researchers become involved in designing AI with enormous cultural impact such as AI-generated art.
This workshop brings together experts on AI development from the NeurIPS and ML community and places them in conversation with experts on human biases and cultural theories from HCI and the social sciences. Although the NeurIPS community plays a vital role in AI development, researchers from outside of machine learning bring key perspectives and methodologies for the empirical study of culture and add nuance to the various ways in which AI and culture intersect.
To this end, the workshop program will focus on the following themes:
- Cultures Shaping AI: How does culture influence assumptions and biases embedded within AI? How do the socio-cultural contexts where AI is developed impact its contours? What harms occur when AI is produced through specific cultural lenses and then deployed and experienced within different global cultural contexts?
- AI Curating Culture: How do AI systems (like recommender systems and language technologies) curate and influence culture? What impact do these technologies have on cultural production and consumption? Which dominant cultures are being centered in the design of these technologies?
- Methods for the AI & Culture Entanglement: What methods and orientations would facilitate understanding the role of culture in AI and AI in culture? What methods can be used to de-center Silicon Valley-centric cultural lenses from AI development to include a broader set of practices? Are there ways that a positive cultural entanglement between society and AI that researchers can promote?
Through a series of inter-disciplinary discussions this workshop will: (a) define the space of problems around AI and culture; (b) stimulate field and theory building around AI and culture; (c) help researchers brainstorm methods and practices for studying AI and culture; and (d) reflect on integrating AI and culture into the research and engineering of AI systems.
Panels
Panel 1: AI Curating Culture
How AI production tools impact culture
Kalika Bali, Microsoft Research
Jennifer Lena, Columbia University
Jonathan Sterne, McGill University
Percy Liang, Stanford University
Panel 2: Cultures of AI
How social, organizational, CS pedagogical cultures influence AI development
Angèle Christin, Stanford University
Mary L. Gray, Microsoft Research
Rumman Chowdhury, Twitter
William Isaac, DeepMind
Schedule
December 9, 2022
| Time (PT) | |
|---|---|
| 09:00-09:15 | Opening |
| Rida Qadri, Fernando Diaz, Alex Hanna, Nick Seaver, Morgan Scheuerman | |
| 09:15-10:30 | Invited Panel 1: AI Curating Culture Moderator: Fernando Diaz |
| Kalika Bali, Microsoft Research Jennifer Lena, Columbia University Jonathan Sterne, McGill University Percy Liang, Stanford University |
|
| 10:30-10:40 | break |
| 10:40-11:55 | Invited Panel 2: Cultures of AI Moderator: Rida Qadri |
| Angèle Christin, Stanford University Mary L. Gray, Microsoft Research Rumman Chowdhury William Isaac, DeepMind |
|
| 11:55-12:30 | break |
| 12:30-13:35 | Contributed Panel 1a: Ecosystems of AI Development Moderator: Morgan Scheuerman |
| Cultural work in ‘AI Ethics and Society’ pedagogy: Some early reflections Maya Indira Ganesh, Rune Nyrup “Moral Imagination:” A Method for Culture Change in Technology Development Amanda McCroskery, Geoff Keeling, Renee Shelby, Sandra Blascovich, Ben Zevenbergen All the tools, none of the motivation: Organizational culture and barriers to responsible AI work Amy K. Heger, Samir Passi, Mihaela Vorvoreanu One Venue, Two Conferences: The Separation of Chinese and American Citation Networks Bingchen Zhao, Yuling Gu, Jessica Zosa Forde, Naomi Saphra |
|
| Contributed Panel 1b: Frameworks of AI/Culture entanglement Moderator: Nick Seaver |
|
| Cultural Incongruencies in Artificial Intelligence Vinodkumar Prabhakaran, Rida Qadri, Ben Hutchinson An Analytics of Culture: Modeling Subjectivity, Scalability, Contextuality, and Temporality Nanne van Noord, Melvin Wevers, Tobias Blanke, Julia Noordegraaf, Marcel Worring The Myth of Culturally Agnostic AI Models Eva Cetinic Recommender systems as cultural technologies: steps to an expanded definition of culture in/and AI Georgina Born |
|
| 13:35-13:45 | break |
| 13:45-14:50 | Contributed Panel 2a: AI and cultural production Moderator: Alex Hanna |
| Redefining Relationships in Music Christian Detweiler, Beth Coleman, Fernando Diaz, Lieke Dom, Jesse Engel, Cheng-Zhi Anna Huang, Larry James, Ethan Manilow, Amanda McCroskery, Kyle Pedersen, Pamela Peter-Agbia, Negar Rostamzadeh, Robert Thomas, Marco Zamarato, Ben Zevenbergen Cultural Re-contextualization of Fairness Research in Language Technologies in India Shaily Bhatt, Sunipa Dev, Partha Talukdar, Shachi Dave, Vinodkumar Prabhakaran Aggregate, Integrate and Align to Embed Everything: A Multi-Modal Framework for Measuring Cultural Dynamics Bhargav Srinivasa Desikan, James Evans When AI Meets the K-Pop culture: A case study of fans’ perception of AI Private Call Eun Jeong Kang, Haesoo Kim, Hyunwoo Kim, Juho Kim |
|
| Contributed Panel 2b: Theorizing AI/Culture entanglement Moderator: Nick Seaver |
|
| Algorithmic Performativity Emanuel Moss Machine Learning as an Archival Science: Narratives behind Artificial Intelligence, Cultural Data, and Archival Remediation Giulia Taurino, David A. Smith Sapir’s Thought-Grooves and Whorf’s Tensors: Reconciling Transformer Architectures with Cultural Anthropology Michael Castelle Training a Model =/≠ Generating Culture: The Meaning of Culture and the Prospect of Artificial Intelligibility Michael W. Raphael, Nga Than |
|
| 14:50-15:00 | break |
| 15:00-15:30 | Closing |
| Rida Qadri, Fernando Diaz, Alex Hanna, Nick Seaver, Morgan Scheuerman |
Accepted Submissions
- Cultural Re-contextualization of Fairness Research in Language Technologies in India, Shaily Bhatt, Sunipa Dev, Partha Talukdar, Shachi Dave, Vinodkumar Prabhakaran
- Recommender systems as cultural technologies: steps to an expanded definition of culture in/and AI, Georgina Born
- Sapir’s Thought-Grooves and Whorf’s Tensors: Reconciling Transformer Architectures with Cultural Anthropology, Michael Castelle
- The Myth of Culturally Agnostic AI Models, Eva Cetinic
- Redefining Relationships in Music, Christian Detweiler, Beth Coleman, Fernando Diaz, Lieke Dom, Jesse Engel, Cheng-Zhi Anna Huang, Larry James, Ethan Manilow, Amanda McCroskery, Kyle Pedersen, Pamela Peter-Agbia, Negar Rostamzadeh, Robert Thomas, Marco Zamarato, Ben Zevenbergen
- Cultural work in ‘AI Ethics and Society’ pedagogy: Some early reflections, Maya Indira Ganesh, Rune Nyrup
- All the tools, none of the motivation: Organizational culture and barriers to responsible AI work, Amy K. Heger, Samir Passi, Mihaela Vorvoreanu
- When AI Meets the K-Pop culture: A case study of fans’ perception of AI Private Call, Eun Jeong Kang, Haesoo Kim, Hyunwoo Kim, Juho Kim
- “Moral Imagination:” A Method for Culture Change in Technology Development, Amanda McCroskery, Geoff Keeling, Renee Shelby, Sandra Blascovich, Ben Zevenbergen
- Algorithmic Performativity, Emanuel Moss
- An Analytics of Culture: Modeling Subjectivity, Scalability, Contextuality, and Temporality, Nanne van Noord, Melvin Wevers, Tobias Blanke, Julia Noordegraaf, Marcel Worring
- Cultural Incongruencies in Artificial Intelligence, Vinodkumar Prabhakaran, Rida Qadri, Ben Hutchinson
- Training a Model =/≠ Generating Culture: The Meaning of Culture and the Prospect of Artificial Intelligibility, Michael W. Raphael, Nga Than
- Aggregate, Integrate and Align to Embed Everything: A Multi-Modal Framework for Measuring Cultural Dynamics, Bhargav Srinivasa Desikan, James Evans
- Machine Learning as an Archival Science: Narratives behind Artificial Intelligence, Cultural Data, and Archival Remediation, Giulia Taurino, David A. Smith
- One Venue, Two Conferences: The Separation of Chinese and American Citation Networks, Bingchen Zhao, Yuling Gu, Jessica Zosa Forde, Naomi Saphra
Organizers
- Rida Qadri (Google)
- Fernando Diaz (Google)
- Alex Hanna (DAIR)
- Nick Seaver (Tufts)
- Morgan Scheuerman (UC Boulder)
Call for position papers
We are looking for 2-3 page position papers that center a provocation on how cultures shape AI design and development or how AI tools shape cultural consumption/production. These provocations can be theoretical, methodological, technical, exploratory or empirical. This workshop is meant to be inter-disciplinary so we welcome submissions of social AND technical papers from scholars of diverse theoretical backgrounds including but not limited to Machine Learning, Computer Science, Anthropology, Sociology, Human-Computer Interaction, and Media Studies. Please make sure your papers are understandable to a broad, cross-disciplinary audience. These submissions are non-archival, will not be part of published proceedings or considered a refereed publication.
Themes include but are not limited to:
- Dominant cultures that shape broader AI solutions
- Cultural assumptions encoded in design decisions of tools like text-to-image models, music production technologies, recommender systems, and language technologies
- Performance of AI technologies in various cultural contexts (e.g. translation systems)
- Harms of deploying culturally inappropriate AI systems in different global cultural contexts
- Disciplinary assumptions that influence technical aspects of machine learning research
- Impact on artists and communities as they use AI for generation, production, and distribution of cultural content
- Current or expected changes in macro-level cultural trends due to AI
We welcome papers focusing on the relationship between culture and social and technical aspects of AI design and development including but not limited to:
- Datasets and training corpora
- Problem formulation
- Imagination of solutions
- Optimization objectives
- Evaluation metrics
- Research and deployment organization
- Other design decisions (e.g. hardware, model parameters)
We take an expansive definition of culture including but not limited to:
- Socio-cultural and geographical context
- Pedagogical/disciplinary cultures
- Institutional cultures
- Creative culture
- Religion
- Broader tech ecosystem
References do not count toward the 2-3 page limit.
Our deadline for submissions is October 6, 2022.
Reviewing
Papers will be reviewed by workshop organizers and selected for:
- Disciplinary depth
- Novelty of perspectives
- Ability to foster cross-disciplinary dialog
Presentation Format
Accepted submissions will be non-archival and presented during the workshop as short lightning presentations.
Dates/Submission portal
All registered participants of NeurIPS can attend the workshop. We may have limited discounted registrations available for participants.
For questions please get in touch with us at: aicultures2022@gmail.com
Submission Deadline: October 6, 2022
Acceptance Notification: October 20, 2022
Workshop Date: December 9, 2022