A collaboration between the Multimedia Analytics research group at IvI and the AI, Culture, and Society research group at the ILLC. All events will take place at Lab42 in the Amsterdam Science Park from 15:30 to 17:00 once a month. Organized by Houda Lamqaddam, Nanne van Noord, Bharath Ganesh, and Gabriel Pereira.

Theme: AI, Visual Culture, and Power

#3 — The creative cultures of AI: 19 February 2025, Lab42, L3.36, 15:30-17:00

Speaker 1: Dr Kim Baraka (VU)

Title: How the performing arts can help us develop more capable interactive robots

Abstract: In this talk, I will discuss the interplay between my background as an interactive robotics researcher and my background as a dance artist. I will discuss concrete ways in which the performing arts, and particularly expertise and knowledge from dance, can inspire, inform, and help test technical advances in interactive robotics, from social robots to physically assistive robots. This will use the following op-ed piece as a starting point https://shorturl.at/yCPHr.

About the Speaker: Kim Baraka is a tenured assistant professor in the Computer Science Department at the Free University (VU) in Amsterdam, and member of the Social AI group. Before joining the VU, he was a postdoctoral fellow in the Socially Intelligent Machines Lab at UT Austin. He holds a dual Ph.D. in Robotics from Carnegie Mellon University (CMU) and the Instituto Superior Técnico in Lisbon, Portugal and a M.S. in Robotics from CMU. His interdisciplinary research, at the cross-roads of human-centered robotics, machine learning, and dance, focuses on enabling robots and humans to teach and learn from each other through situated social interactions. He is part of the Hybrid Intelligence Center, which fundamentally researches hybrid human-AI systems, with applications to healthcare and education. As a professionally trained contemporary dancer, he is particularly interested in new frontiers in robotics that draw inspiration from the performing arts.

Speaker 2: Dr Rosa Wevers (Minerva Academy, Groningen)

Title: Algorithmic Surveillance: A curatorial Inquiry

Abstract: This presentation explores the critical potential of curatorial inquiry in the realm of algorithmic surveillance. I reflect on the exhibition ‘Face Value’ (2021), a research-exhibition (Sheikh 2015) I curated in collaboration with IMPAKT [Centre for Media Culture]. Through a series of artistic interventions, ‘Face Value’ sought to defamiliarize the normalized technologies of algorithmic surveillance, revealing their hidden politics. By creating these defamiliarizations, the exhibition aimed to engage visitors not just in sense-making but also in sensing surveillance, ultimately offering tools for them to become more responsive to it. In doing so, the research-exhibition aimed to foster a critical surveillance spectatorship.

About the Speaker: Dr. Rosa Wevers works as Senior Researcher Art & Technology at Research Centre Art & Society (part of Minerva Art Academy in Groningen). She is co-curator at Noorderlicht (platform for lens-based media), and host of the podcast Kunstmatig. In her research and curatorial practice, Rosa explores the relations between art, technology and society. Rosa obtained her PhD at Utrecht University. Her project explored how contemporary art exhibitions confront visitors with critical perspectives on surveillance and engage them in strategies of resistance. As part of her PhD, Rosa curated the exhibition ‘Face Value’ in 2021, in collaboration with 2021 IMPAKT [Center for Media Culture] and NFF.

Past seminars:

#2 — AI in Film, From Production to Analysis: 30 January 2025, 15:30–17:00, Lab42, Room L3.36, Amsterdam Science Park

This seminar will delve into the role of AI in film, exploring its application in both creative production and critical study.

Invited Speakers:

Abstract: The integration of artificial intelligence (AI) into audiovisual production represents a transformative shift in the creation and analysis of cinematic narratives. At Artefacto, we explore the intersection of AI and filmmaking through research and production that combines technological innovation with critical reflection. This presentation examines case studies from Artefacto’s ongoing projects, including OpenDocs, an initiative providing documentarians with AI tools to expand the creative and technical possibilities of storytelling. Drawing from a qualitative and quantitative analysis of over 300 short films and AI-driven works, we highlight recurring trends in the application of generative AI. These include iterative workflows for narrative development, character design, and visual production, alongside insights into the challenges of ethical authorship and creative agency. Through the analysis of large-scale data and multiple case studies from Artefacto’s ongoing projects, including short films and experimental works, we aim to critically assess how AI is fostering new forms of audiovisual representation. These case studies illustrate the integration of traditional cinematic methodologies with cutting-edge AI technologies, showcasing how Artefacto innovates and redefines workflows, narrative structures, and visual aesthetics while critically interrogating the evolving landscape of filmmaking.

About the speaker: Jorge Caballero is a versatile creator and researcher working at the crossroads of technology, cinema, and social impact. He holds a Ph.D. in Communication (Cinema and Artificial Intelligence) from Pompeu Fabra University and is an Associate Professor at both Pompeu Fabra University and the Autonomous University of Barcelona. In these roles, he leads workshops that delve into the intersection of technology, filmmaking, and social transformation. He also co-directs the Postgraduate Program in Film Algorithms and AI at UAB, shaping the next generation of creators in the era of artificial intelligence. Caballero is the co-founder of GusanoFilms and Artefacto, two pioneering production companies whose works have premiered at renowned festivals such as Rotterdam, SXSW, and HotDocs. A two-time recipient of Colombia’s National Documentary Award, his projects span virtual reality, AI-driven storytelling, and collaborations with institutions like the Sundance Institute and Tribeca Film Institute. His latest book examines artificial imagination, highlighting its potential to revolutionize storytelling while emphasizing inclusivity, accessibility, and ethical innovation in the field.

Title: Identifying Nudity in National Cinemas: Potentials and Pitfalls of Computer Vision

Abstract: Recent advancements in computer vision have piqued the interest of film scholars given their potential to test hypotheses about large film corpora, such as the film productions of entire countries through the years. As a case study, we focus on the reputation that certain national cinemas (in particular, the Netherlands, Sweden, and Denmark) have for containing more frequent displays of explicit nudity. In order to test whether this reputation is warranted, we have applied computer vision methods in combination with close readings of a sample of images from a swathe of Dutch, Scandinavian, and US (known for their relative lack of nudity) films. We take as our corpus films shown to Amsterdam audiences in the sample years of 1952, 1962, 1972, and 1982. Our results indicate that while current open models are prone to errors, particularly in the shape of false positives, they can provide convincing evidence of trends and indicate that the Dutch and Scandinavian reputations for nudity are warranted. Furthermore, the process of engaging with such models forced us to face just how subjective the definition of nudity is, since we had relatively high inter-coder disagreement during manual coding. Ultimately, our study underscores the necessity of reflexive and critical scholarship that is interdisciplinary and blends digital methods with humanistic interpretation. Amidst the AI hype, it’s crucial to recognize the complexities inherent in historical analysis and engage in nuanced discussions about the role of technology in shaping our understanding of cultural phenomena.

About the speaker: Isadora Campregher Paiva is a lecturer in Film Studies at the University of Amsterdam (UvA), a PhD candidate at UvA’s Amsterdam School for Heritage, Memory and Material Culture and a research associate in the project Social Sciences & Humanities Open Cloud (SSHOC-NL). She has a Master’s in Sociology from the Universidade Federal do Rio Grande do Sul (UFRGS) and an MA in Audiovisual and Cinema Studies from the Goethe University Frankfurt, where she also worked as a lecturer and a research associate on the project “Digital Cinema-Hub”. Her research focuses on the application of digital methods to the study of film historical questions. This particular presentation is the result of work undertaken at UvA’s CREATE Lab together with a team of collaborators.

#1 — Inaugural Seminar: 20 November 2024, Lab42, L3.33, 15:30-17:00

Speaker 1: Dr Abdallah El Ali (CWI, UU)

Abstract: As AI-generated information proliferates online and becomes increasingly indistinguishable from human-sourced information, it becomes critical to understand how we trust and label such content, especially should such information be inaccurate. For example, misleading labels or unclear sourcing can result in misinformation and poor health decisions, even if (honest) AI system disclosures are increasingly mandated by regulations (e.g., European AI Act). In this talk, I will highlight the importance of considering ‘what’, ‘how’, and ‘when’ AI disclosures are given within two domains, news media content and personal health information. I will show early results on how AI transparency disclosures can impact people’s perceived trust, behavior, and physiological indicators. I will conclude by highlighting our research efforts for creating responsible human-AI interactions at the user interface level, specifically the potential for developing intelligent user interfaces that can infer if and when to apply AI disclosures using sensed behavioral and physiological data.

About the Speaker: Abdallah “Abdo” El Ali is a research scientist at Centrum Wiskunde & Informatica, and part-time Assistant Professor at Utrecht University. He leads the research areas on Affective Interactive Systems and Trustworthy Human-AI Interaction, where he combines advances in sensing and actuation, eXtended Reality, and Artificial Intelligence to augment human cognitive, affective, and social interactions. He is also affiliated with the AI, Media, and Democracy Lab, where he leads Human-AI Interaction research focusing on AI transparency in media. He serves as Chair of CHI Nederland (CHI NL), the ACM SIGCHI Netherlands Chapter. Website: https://abdoelali.com/

Speaker 2: Prof Sabine Niederer (HvA)

Title: From prompt engineering to prompt design. Prompting generative visual AI for [biodiversity]

Abstract: This presentation briefly outlines a conceptual and practical agenda for doing research with generative visual AI that moves from prompt engineering towards prompt design. It illustratesone particular prompt design strategy -that of ‘ambiguous prompting’- with a small case study on bias in AI-generated representations of biodiversity.

About the Speaker: Dr. Sabine Niederer is Professor of Visual Methodologies at the Amsterdam University of Applied Sciences (www.visualmethodologies.org), head of ARIAS – platform for artistic research (http://arias.amsterdam) and co-coordinator of the Digital Methods Initiative at the University of Amsterdam (www.digitalmethods.net). Together with Gabriele Colombo, she recently published the book Visual Methods for Digital Research (2024).