About

Watching Machines is a three-day exhibition and public programme on algorithmic vision: facial recognition, biometric remote identification and behavioural-analysis systems (emotions, gait, movement) are increasingly woven into the infrastructure of public space. Bringing together four years of work from the ERC-funded Security Vision project, it gathers films, installations, and interactive data works by Cyan Bae, Francesco Ragazzi, and Ruben van de Ven, alongside a public programme of conversations with policy makers, artists, researchers, and activists.

The exhibition asks a set of plain questions. Where are these systems deployed, and by whom? On what data are they trained, and through what kinds of human labour? What do they get wrong, and what does that error reveal? And what would it mean to look back at the machines that are looking at us?

Free entry. All welcome.

Works on view

Mapping the field of security vision

Two collective interactive datavisualisations

The Security Vision project has spent four years tracing where algorithmic surveillance systems are deployed across the world, who builds them, and who buys them. The result is a database of 837 deployments across 2,181 institutions in 131 countries. The team also spend longer periods of time talking to specialists in the field, and recording these interviews through an innovative audio-diagramming technique. These two interactive works open that database to the public, turning a research dataset into something that can be navigated, queried, and read. They are an attempt to answer a question that sounds simple but is surprisingly hard to answer with precision: where, exactly, is algorithmic vision happening, and what does it mean?

Welcome to Set

Cyan Bae, 2026, experimental short documentary. Premiered at International Film Festival Rotterdam 2026. International premiere and Tënk Opening Scenes Award at Visions du Réel 2026.

What happens on the sets where AI emotion-recognition systems are trained? Cyan Bae reenacts the scientific procedures by which "ground truth" is generated for machines designed to read human faces: the performances, interviews, and ratings through which raw human expression is processed into machine-readable affect. The light, almost playful tone of these procedures sits in deliberate tension with what they ultimately serve, namely the deployment of algorithmic systems into the field of security. Welcome to Set is the first part of a trilogy on emotional AI and deception detection.

Work in progress shorts

Cyan Bae, two upcoming shorts (work in progress)

Excerpts from the second and third parts of Cyan's trilogy. The films ask how queer and feminist perspectives might disrupt the algorithmic systems that render some bodies suspicious while letting others pass, and reimagine our relationship to them through collaborative filmmaking.

An Algorithm of Violence

Francesco Ragazzi, (work-in-progress preview)

A documentary built around Arsen, a coder in Yerevan developing a violence detection algorithm. The film follows the everyday labour of teaching a machine to recognise a punch, a shove, a fight, and uses that intimate ground to ask larger questions: what do we mean when we ask a system to recognise violence? Whose violence becomes visible, whose disappears, and what aesthetic and ethical decisions are folded into the gesture of training?

Perplexity

Ruben van de Ven, media installation. Previously presented at NEXT LEVEL, Dortmund.

A surveillance camera, a computer, and a set of laser projectors. The system continuously trains itself on the movement of people through the space, predicts their most likely paths, and projects those predicted paths as glowing traces on the floor in front of their feet. Perplexity reverses the logic of conventional surveillance: instead of detecting the anomalous, it simulates the normal. Deviation is what occurs when reality stops matching the simulation, and the algorithm's error becomes a measure of suspicion. Visitors are invited, gently, to disrupt their own routines.

Programme

Wednesday 18 June evening: Opening drinks and exhibition preview.

Thursday 19 June: Public programme: a half-day of conversations with policy makers, artists, researchers, and activists on the politics of algorithmic vision. Full programme to be announced.

Friday 20 June, all day Exhibition open to the public.

Practical info

Venue The Grey Space in the Middle, Paviljoensgracht 20–24, 2512 BP The Hague

Accessibility The Grey Space is wheelchair-accessible. All spaces are reachable by wheelchair and mobility scooter; the venue has a lift and accessible toilets.

Tickets Free entry. RSVP recommended for the opening on the 18th and the public programme on 19th to make sure you can get in!

About the project

Watching Machines presents work from Security Vision: The Algorithmic Security Politics of Computer Vision, an ERC Consolidator Grant project (2020–2025) led by Francesco Ragazzi at the Institute of Political Science, Leiden University. The project investigates the political, social, and aesthetic dimensions of computer vision systems used in security, combining critical security studies with practice-based and multimodal research methods.

Artists & researchers Cyan Bae · Francesco Ragazzi · Ruben van de Ven

Supported by European Research Council · Leiden University · The Grey Space in the Middle · ReCNTR