Book design: Uncovering Algorithms
Role
Company
Timeline
Book design: Uncovering Algorithms
Printed publication
2025
OVERVIEW
Uncovering Algorithms: Conversations on the Impact of AI is a printed book built around interviews with researchers, artists, technologists, and critical thinkers. Conducted and edited by an award-winning AI expert Natalia Domagala.
The chapters examine how AI affects labour, democracy, creativity, climate, discrimination, Indigenous data sovereignty, surveillance, and everyday digital life.
I co-designed the book, focusing on its visual concept, illustration system, image treatment, cover, dust jacket, and editorial direction. Final typesetting was handled by Kuba Bogacki.
Book is available on Metalabel and Antenne Books.
The aim was to create a book that felt accessible without making AI appear frictionless, neutral, or detached from the people, labour, infrastructure, and environments affected by it.
CHALLENGE
Many systems discussed in the book are difficult to see directly. Data extraction, outsourced labelling, cloud infrastructure, automated classification, and surveillance are often hidden behind interfaces and the language of frictionless technology.
The design needed to make these systems more visible while treating subjects such as exploitation, discrimination, environmental harm, and digital colonialism with care. It also had to remain readable and approachable without reducing complex political issues to decoration.
How do you design a book about AI that is readable and accessible, while still respecting the harm, complexity, and political weight of the subject?
DISCOVERY
The visual research began with two questions: what can a reader recognise immediately, and what can create the uncanny feeling of being watched?
Facial recognition became one of the strongest entry points. It connects abstract technology with the body, identity, privacy, and the feeling of being watched. It also reveals the logic of machine classification: a face becomes a grid, a set of points, a category, or a probability.
Emotion recognition extended this idea. Complex, contextual, and culturally dependent expressions are reduced to a small number of fixed labels. This reduction of human ambiguity into technical outputs became one of the main visual threads of the book.
Research-based imagery was used carefully to show this process without treating real harm as visual material. The intention was to expose the violence of simplification: how systems make people readable to machines by removing uncertainty, context, and difference.
Other references came from familiar digital experiences. CAPTCHA, privacy policies, checkboxes, system messages, and old software environments connected large technical systems with ordinary actions. Readers might not understand the mechanics of machine learning, but they recognise the experience of agreeing to unread terms or selecting traffic lights to prove they are human.
The direction moved away from generic AI imagery such as robots, glowing networks, blue gradients, and futuristic surfaces. The book needed to feel grounded in systems that already shape daily life.
SOLUTION
☞ Turn recognition into unease
Facial recognition, emotion classification, CAPTCHA, consent checkboxes, and system messages were used as familiar entry points. Readers already knew these visual languages, but placed in a new context, they became unsettling.
Research-based imagery showed how faces and expressions are reduced to points, grids, labels, and probabilities. The aim was to expose the distance between human experience and the simplified categories used by technical systems.
☞ Make digital infrastructure feel physical
ASCII graphics, technical diagrams, gears, and mechanical illustrations challenged the idea of AI as abstract or weightless.
Concepts such as “the cloud” were brought back to their material foundations: servers, energy, land, extraction, and labour. Human figures placed inside technical systems suggested people being measured, processed, and absorbed into machinery.
☞ Build the concept into the book itself
The visual language extended beyond individual illustrations.
The dust jacket used a dense glitch collage, while its inside contained terms and conditions in very small text. Underneath, the cover was almost empty.
Removing the jacket became part of the concept: moving from visual overload to the quieter structures hidden beneath it.
Across the publication, research imagery, ASCII, interface references, distortion, and abstract chapter openings formed one coherent system around surveillance, classification, hidden labour, consent, and power.
REFLECTION
Sometimes the work is about making invisible structures visible enough to question them. Enough to notice the labour, extraction, classification, energy use, and politics behind the interface.
It also taught me how careful design has to be when the content deals with real harm. Exploitation, discrimination, surveillance, digital colonialism, environmental damage, and hidden labour cannot become decoration.
