Back to contents

Design against the machine

This course at FH Potsdam explored the intersection of design and generative AI, challenging students to build experimental websites that critically engage with AI-generated content. Legal, ethical, and creative questions were central: who does AI-generated work actually belong to, and what does it mean when machines can replicate human creativity at scale?

A distressed halftone landing page titled ‘Welcome to the biggest image-database in the world’ with an EXPLORE button.
1.1 The world’s biggest image database. You just can’t see any of it.
An AI-generated collage of a globe, screens and figures in saturated colour, surrounded by hand-drawn icons.
1.2 An early version, before the aesthetic caught up with the concept.
A surreal AI image of a robed figure kneeling before a glowing globe, surrounded by tiny pixel icons and a Submit button.
1.3 Click the symbols in the right order. The correct order does not exist.
A stormy AI painting reduced to a ‘Solve the CAPTCHA’ challenge with numbered points and pixel icons.
1.4 The image used to be something.
A nine-tile captcha grid overlaid on a blurred ‘Welcome’ hero, asking to select subway entrances.
1.5 Select all squares with subway entrances in a painting from a century without subways.

Cryptic Captchas presents itself as a vast image archive, but confronts visitors with captchas that grow increasingly absurd and unresolvable. The more you interact, the less certain you become of what is content, what is generated, and where the boundary lies. It probes how AI reshapes our relationship to digital images and the act of seeing itself.

The core challenge was building something that genuinely defies both human and machine interpretation. The site was built in HTML, CSS, and JavaScript, using AI-generated imagery as raw material. Then I deliberately obscured, fragmented, and recombined it to create unresolvable verification tasks. The captchas escalate in difficulty across sessions, rewarding persistence with deeper disorientation.