Project Details
Watershed is an online art project that explores the algorithmic gaze in an era of digital news. It employs a computational process to dismantle and reassemble the visual torrent of daily media, inviting users to navigate a world built from the fragments of our shared informational landscape. It is an experiment in seeing, questioning not just what we see in the news, but how we are conditioned to see it.
Each day, Watershed collects images from international news websites and subjects them to a watershed segmentation algorithm. This process dismantles the images into fragments by “flooding” a grayscale version of an image from its darkest points to its lightest, just as rainwater would fill a real-world mountainous landscape from its lowest valleys. In doing so, it performs a form of algorithmic seeing quite different from that of a human—one that operates without narrative, hierarchy, or significance.
The system does not distinguish between what a photojournalist might frame as a central subject and what might appear incidental. A protestor’s sign, a celebrity’s face, a patch of sky, or a piece of rubble—all are treated as equivalent grayscale image data to be segmented and archived.
Online, these fragments are assembled pseudo-randomly into a navigable three-dimensional space. The goal is not to retell the news but to reorganise its visual material according to a logic other than the ones to which we have become habituated. When compositional and editorial hierarchies are systematically dismantled, how does the perceived importance of visual information change? This reordering not only alters the image itself but also reshapes how it must be encountered.
To experience this material, users navigate a digital environment rather than scroll through a linear feed. The familiar rhythm of news consumption—fast, frictionless, designed for scanning—is replaced by movement through a spatial terrain. Seeing becomes a form of exploration.
The aesthetic of this space is deliberately glitchy and non-photorealistic. As a user zooms in on a fragment, its pixel structure reveals itself, breaking the illusion of seamlessness. Deconstructed 2D news images drift across a constructed 3D landscape, never pretending to be windows onto reality. This makes the artifice of the medium explicit: images are data, not evidence.
By slowing down the act of viewing, the project alters the relationship between attention and imagery. Standard interfaces are optimised for speed, reducing engagement to rapid recognition. Here, engagement is physical and deliberate. The contrast is not presented as a solution but as a tool of critique—making the invisible conventions of ordinary web interfaces visible by inverting them.
The project arises from a media environment where trust in both institutions and images is unsettled. With deepfakes and decentralised information sources, the photograph’s credibility as evidence is under renewed scrutiny. Yet, as critics such as Allan Sekula have argued, photography has never been a transparent medium; it has always been a construction.
What has changed is scale and automation. Today, images are filtered, analysed, and manipulated by systems we rarely notice or understand. This forms a new layer of mediation—one that is constant, pervasive, and almost invisible. It is this everyday, digital infra-ordinary that Watershed makes tangible.
This invisibility operates on several levels. We are shaped by the grammar of the frame, where composition and colour psychology direct our gaze and mood. We are guided by digital defaults, where smartphone cameras enhance, correct, or beautify before we even see the result. We are steered by algorithms that curate which images reach us. And we interpret everything through cultural repertoires of symbols and social cues. By dismantling these logics and scrambling their effects, the project forces us to confront how much of our seeing is already conditioned.
The project also speaks to a transitional moment. The centralised gatekeeping of 20th-century journalism has given way to fragmented, unstable formats. We are collectively searching for new grammars of public discourse. Rather than proposing a return to older models, Watershed treats this uncertainty as an opportunity: a chance to experiment with new arrangements of familiar material, to test how different structures shape perception. It suggests that the media of the future may not converge into a single stable format, but instead coexist as a patchwork—some fostering clarity, others producing estrangement.
This approach shares a lineage with experimental practices such as Surrealist automatic writing and collage. By establishing rules and constraints—writing without interruption, juxtaposing unrelated images—the Surrealists sought to bypass rational control and provoke unexpected associations. In a parallel way, Watershed treats the algorithm as a collaborator. The code sets the procedure, but the visual outcomes remain emergent and unpredictable.
There is a creative pleasure in this surrender of control: in letting a system run and being surprised by the strange, poetic, or nonsensical landscapes that result. Like collage, the work is not about arriving at a definitive meaning but about generating provisional configurations that open new ways of seeing.
At its heart, the project asks: what meaning remains in news images once they are stripped of context, headlines, and captions, reduced to fragments of colour and form? The result is a fleeting, fractured archive of the collective present—an algorithmic shadow of daily events. It invites reflection on both what we choose to remember and what the machine, indifferent to significance, simply sees.
