FACES AND MACHINE VISION: PUBLIC CONVERSATIONS ON FACIAL AI AND SOCIETY
4 November 2026, Building 4, Room A302
CIES–Iscte, University Institute of Lisbon (IUL), Lisbon [Avenida das Forças Armadas, 40, 1649-026, Lisboa]
Facial recognition and other forms of facial AI are increasingly used across policing and border control, smartphones and digital platforms, healthcare and education, and a growing range of everyday environments. In this context, facial AI refers broadly to AI-enabled systems that detect, analyse, or infer information from faces, such as identification/verification, categorisation, emotion/affect inference, attention and behaviour analytics, and other predictive or profiling applications.
This meeting brings together a diverse international group of social sciences and humanities experts–spanning STS, semiotics, art and cultural studies, policy, design, visual culture, and surveillance studies–to discuss a timely and urgent question: what is at stake when facial AI makes the face a key interface for decision-making, governance, and everyday life?
This event explores how machine vision reshapes what counts as being “seen” (and what becomes invisible); how design choices embed power and assumptions; how masks, avatars, and algorithmic “faciality” rework identity and recognition; how surveillance systems reproduce racial and gender inequalities; how computational approaches can help map public controversies; how histories and cultural narratives of AI influence what becomes thinkable or inevitable; and what “responsible” governance could look like in practice.
Guiding questions include: What does it mean to be “seen” by a machine? Who defines what is visible, measurable, and actionable? Which inequalities are amplified and under what conditions? How do surveillance systems become normalised within institutions (schools, campuses, workplaces)? What does a feminist/decolonial approach change in how risk, safety, and “user needs” are defined? What are the alternatives to facial recognition beyond a simple “ban vs allow” framing? How can artistic and creative practices help publics imagine and debate different futures?
PROGRAMME
8:30 – Reception to participants
9:00 – Opening (Patrícia Ávila, Director of CIES-Iscte; Helena Machado, PI of fAIces project)
9:30 – 10:45 Session I
Seeing & Being Seen (Moderator: Guya Accornero, CIES-Iscte)
Beyond the Face: Surveillance as a Tool of Domination on U.S. University Campuses – Torin Monhan, University of North Carolina at Chapel Hill, United States of America
Facial Recognition: Gatekeeper of Citizenship and Citizenship Resources – Aaron Tucker, Memorial University of Newfoundland and Labrador, Canada
Uncanny Faces: Data Extraction and Dividual Images – Fernanda Glória Bruno, Federal University of Rio de Janeiro (UFRJ), Brazil
10:45 – 11:00 Coffee break
11.00 – 12:15 Session II
Controversies & Governance Futures (Moderator: Jussara Rowland, ICS, University of Lisbon)
Intersectionality, Bias and Artificial Intelligence: Technical Fixes, Wicked Problems and Framing of Policy Prescriptions – Inga Ulnicane, University of Cambridge, United Kingdom
The Laws of the Face – Massimo Leone, University of Turin, Bruno Kessler Foundation, Italy
Another Form of Biometrics: Genetics and the Politics of Security – Joëlle Vailly, National Centre for Scientific Research CNRS, IRIS, France
(Lunch)
14:30 – 15:00
Artistic Performance (Introduced by Susana de Noronha, CIES-Iscte)
trAIts – L’Homme Moyen, by Pedro Alves da Veiga, CIAC – Research Centre in Arts and Communication, Universidade Aberta, Portugal
15:00 – 16:15 Session III
Faces, Masks & Inequality (Moderator: Frederico Duarte, Faculty of Fine Arts, University of Lisbon)
Invisible by Design: How Algorithmic Surveillance Systems Fail, Control, and Silence Women – Miren Gutierrez, University of Deusto, Spain
Politics of Design – Per/Forming Critique – Michelle Christensen, Berlin University of the Arts/ Weizenbaum Institute; Technische Universität Berlin, Germany
On Machinic Ways of Seeing and the Composite Form – Lila Lee-Morrison, Lund University, Sweden; Irwin S. Chanin School of Architecture, Cooper Union, United States of America
16:15 – 16:30 Coffee break
16:30 – 17:30
Closing roundtable (Moderators: João Santos, CIES-Iscte & Susana Silva, ICS-Uminho)
SPEAKERS & ABSTRACTS
Torin Monahan (PhD) University of North Carolina at Chapel Hill, United States of America.
Beyond the Face: Surveillance as a Tool of Domination on U.S. University Campuses
In this presentation, I explore ways that identification-based surveillance systems, such as facial recognition, embed the people targeted by them within larger relations of domination. Drawing on a study of the policing of student activists in the United States, I show how while surveillance exposure may enable discrete moments of control, its deeper threat lies in processes of distanciation that open activists up to future forms of punishment. For instance, the identification of student activists protesting genocide in Gaza has led in some cases to their later abduction and detention by Immigration and Customs Enforcement (ICE) agents, visa revocations, doxing by Zionist groups, suspensions and expulsions, scholarship termination, and employment restrictions. Importantly, students face differential forms of vulnerability based on their nationality, race, and immigration status. The unknowability of whether or how one will encounter reprisal for exercising free-speech rights injects an element of fear or terror into protest activities. Notwithstanding creative exercises of resistance by activists, the situation catalyzes chilling effects and conditions social conformity, which ultimately support right-wing ideologies on university campuses and beyond.
Aaron Tucker (PhD) Memorial University of Newfoundland and Labrador, Canada.
Facial Recognition: Gatekeeper of Citizenship and Citizenship Resources
As initially detailed by 404 Media, and expanded upon by the BBC in January 2026, United States Immigration and Customs Enforcement has been using the facial recognition technology (FRT) Mobile Fortify as part of its latest set of tactics stemming from the Trump administration’s deeply harmful immigration crackdown. Mobile Fortify, with a database of 200 million photos, leverages facial images captured as part of the Customs and Border Protection’s practice of capturing individuals’ faces as they enter and exit the United States. This presentation draws from the forthcoming monograph Flexible Faces (McGill Queen’s Press, Nov. 2026) in order to ground such contemporary instances of FRTs in a historical trajectory beginning with the technology’s earliest forms in the 19th and early 20th century. This arc has entrenched facial recognition as a tactic rooted in eugenics and operating under the rationales of national security and health, where FRTs operate as a tool crucially shaped by forms of “desirable” and “undesirable” citizenship, despite the well-documented racial and gendered biases of the technology.
Fernanda Glória Bruno (PhD) Federal University of Rio de Janeiro (UFRJ), Brazil.
Uncanny Faces: Data Extraction and Dividual Images
This talk examines the trajectory of facial images within the digital ecosystem, focusing on the constitution of the face as a site of data extraction and on the use of facial datasets to train generative artificial intelligence systems for the production of synthetic faces. It aims to analyze the status of these faces as simultaneously strange and familiar, as well as the dividual condition of generative images.
Inga Ulnicane (PhD) University of Cambridge, United Kingdom.
Intersectionality, Bias and Artificial Intelligence: Technical Fixes, Wicked Problems and Framing of Policy Prescriptions
Artificial Intelligence has often been presented as a neutral and objective tool that will help to eliminate or at least reduce human bias. However, evidence suggest that rather than mitigating human bias AI is exacerbating and amplifying it as can be seen, for example, in facial recognition systems that work best on white male faces and worst on black female faces or sentencing algorithms that overestimate potential of reoffend of black defendants while underestimating chances of white defendants to reoffend. Against this background, this talk draws on analysis of AI policy documents examining how they frame intersectional bias, its causes, impacts, and recommendations for action. The analysis reveals that rather than treating AI as a quick technical fix to gender, racial, socio-economic and other types of biases and focusing on increasing numbers of women and minorities in tech sector, it is important to develop holistic approaches that treat bias as a complex and uncertain ‘wicked problem’ and address related cultural, structural and power issues as well.
Massimo Leone (PhD) University of Turin, Italy; Bruno Kessler Foundation, Trento, Italy.
The Laws of the Face
This paper reframes facial recognition beyond the sterile “ban vs allow” dichotomy by situating it within a plural field of legal semiotics, governance models, and emerging computational paradigms. Rather than asking whether FRT should exist, it investigates how societies construct the face as a juridical object across different regulatory grammars: sensitive biometric data, prohibited technological actant, context-dependent policing practice, or federated regulatory patchwork. Through comparative analysis (EU, China, UK, North America, Canada, India), the paper shows that governance reflects deeper semiotic ideologies of visibility, authority, and social trust. Moving beyond current debates, it examines two transformative trajectories: quantum-inspired computation and agentic AI. While quantum FRT remains largely speculative, agentic AI introduces autonomous systems capable not only of recognizing but interpreting, predicting, and acting upon faces as relational interfaces. This shift raises new legal and sociocultural challenges: delegation of perception, algorithmic agency, anticipatory profiling, and the reconfiguration of accountability. The paper proposes layered governance combining enforceable opacity zones, human-in-the-loop architectures, and socio-legal frameworks preserving personhood under agentic mediation.
Joëlle Vailly (PhD) National Centre for Scientific Research (CNRS), France.
Another Form of Biometrics: Genetics and the Politics of Security
My presentation will focus on a biometric technique based on DNA analysis that raises issues often overlapping with those of facial recognition. My research aims to analyze how the suspect and her/his body is constituted as objects of genetic knowledge, as a vessel of evidence and proof, and as an instrument of the government of security. Based on a study conducted in France over a period of eight years, it seeks not only to examine the construction of suspects enabled by legislation and public policy, but also to grasp the rationale underpinning this construction, its forms and its effects, through the discourses and practices of agents of the state, managers of biotechnology companies, and of the individuals to whom these methods are applied. It shows that DNA is not only a tracker of suspicion, but also a marker of contemporary social developments.
Miren Gutierrez (PhD) University of Deusto, Spain.
Invisible by Design: How Algorithmic Surveillance Systems Fail, Control, and Silence Women
Algorithmic surveillance technologies — from facial recognition to predictive policing and welfare fraud detection — are routinely analyzed through the lens of racial bias, yet their gendered harms remain under-examined and under-documented. Drawing on a comparative analysis of algorithmic-based decision systems across policing, healthcare, social services, and border control, this talk maps the ways in which these technologies simultaneously render women invisible (through training data gaps, under-reporting, and research lacunae) and hyper-visible (through systems designed to monitor, discipline, and control their bodies and behaviours). Cases such as Iran’s facial recognition enforcement of hijab laws, VioGen’s failure to protect domestic violence victims in Spain, and the gendered biases embedded in machine translation at borders illustrate how machine decision-making becomes a site of gendered state violence. The talk argues that a feminist analysis of algorithmic surveillance must move beyond bias audits toward a structural critique of whose safety these systems are truly designed to serve.
Michelle Christensen (PhD) Berlin University of the Arts/ Weizenbaum Institute; Technische Universität Berlin, Germany.
Politics of Design – Per/Forming Critique
As artificial intelligence finds its way into the mundanity of everyday life, scanning and categorising us in order to provide the highest level of comfort and efficiency, the systems that surround us are increasingly mediating our knowledge, actions and behaviours. Neatly presented and packaged to relieve our regular routines, these technologies colonise the internet with perceptions, practices and probabilities that include some, and exclude others. Intersectional gender bias in AI produces precarious probabilities, and these implicit and explicit deeds of design culminate not just in discursive matters of grave concern, but in the manufacturing of im/possibilities into perceiving histories and presents. This talk explores how design and artistic research can be applied to imagine and prototype alternative technological futures using open-source tools and interdisciplinary, open science approaches. Drawing on research into prototyping offline feminist AI systems and interdisciplinary student projects, it presents experiments that demonstrate how performative practice can interrogate AI systems while opening space for more inclusive and participatory futures.
Lila Lee-Morrison (PhD) Lund University, Sweden; Irwin S. Chanin School of Architecture, Cooper Union, United States of America.
On Machinic Ways of Seeing and the Composite Form
This talk draws on a section of the book, Portraits of Automated Facial Recognition: On Machinic Ways of Seeing the Face that examines the composite form as a key visual logic of algorithmic perception. Through tracing an epistemic history of composite representation in histories of facial portraiture, it addresses how it has operated as a form of abstraction which mediates between the general and the particular. In referencing composite portraits in contexts of both art and philosophy, I look at how composite logic can be generative in a way that expands beyond a dualistic approach towards representing the general and particular and instead draw on notions of community and utopic potentials of seeing anew. In addressing the composite form as a central visual logic of algorithmic perception and image generation, this talk contributes to ongoing analyses on the aesthetics of AI mediation.
ARTIST & SYNOPSIS
Pedro Alves da Veiga (PhD) CIAC – Research Centre in Arts and Communication, Universidade Aberta, Portugal
trAIts: L’Homme Moyen
trAIts – L’Homme Moyen is an interactive installation-performance that stages a device promising to identify who a person is from their facial biometry. It delivers on that promise with the taxonomic confidence of nineteenth-century anthropometry, criminal anthropology and composite portraiture, and of the “physiognomic AI” that reanimates them today. Within the room the work does not present itself as satire or as critique; it behaves as a serious identification system. Its argument is not that the machine errs, but that it returns the truth — its own design and sample already contained. Each participant receives an individual verdict in a clinical register; after a set number of interactions, the room receives a collective report — the “average man” the system has itself produced.