Immediate help clickandstop.ch

One face. Two narratives.

The same person in two completely different stories. How does our view of people change when their image is manipulated?

Image
Image
Image
Image
Image
Image
Image
Mann mit verschränkten Armen vor einem See mit Wald und Bergen im Hintergrund.
Mann mit lockigem Haar und Bart in einem schmutzigen weißen Hemd vor einer Größenmesswand.
The left picture

The left image shows Ramin, known from SRF Zwei am Morgen: authentic, sporty, in nature.

The right picture

The right one is an AI fake. His face was turned into a criminalising mugshot.

The danger: AI reinforces racist and discriminatory patterns. People with a migration background or people of colour are more quickly pushed into problematic narratives.

AI artist Basil Stücheli shows how a deceptively realistic fake photo is created from the original image from the Ringier archive by Stefan Bohrer – and why this can be dangerous.

My work is defined by authenticity, fleeting moments, and the responsibility to capture reality as genuinely as possible. This is precisely where I see the primary danger of AI: it increasingly blurs the line between documentation and fiction.

Fotograf Stefan Bohrer

How does this image manipulation affect you?

Stigmatisation

One click can turn someone from a friend into a suspect.

Bias exposed

AI systems adopt discriminatory patterns from training data.

Real-world consequences

Such depictions can jeopardize jobs, relationships, or even the safety of the individual concerned.

Reinforce prejudices

A single image can quickly lead to a snap judgment, and AI-generated fakes have the potential to further entrench existing stereotypes.

#OnlyFacts

Gruppe von Models mit Spiegelreflexionseffekt in einem loftartigen Raum mit Backsteinwand.

Facial recognition has up to a 35× higher error rate for people of colour than for white people.

(Source: ACLU Minnesota, 2018)

Gruppe vielfältiger Menschen in variierenden weißen Outfits in modernem Raum mit großen Fenstern.

AI models reproduce discrimination unnoticed — the bias is in the data.

(Source: AI Now Institute, Landscape Report 2023)

Menschen unterschiedlicher Herkunft in weißer Kleidung stehen sich in zwei Reihen gegenüber.

People judge the same person completely differently depending on the image context.

(Source: Journal of Experimental Psychology, 2020)

Mehrere Personen in weißen Anzügen in einem Raum mit Spiegeln, die Reflexionen erzeugen.

49% of the Swiss population are concerned that AI could lead to discrimination.

(Source: Digital Barometer, 2024)

#ClickWithCare - Your bias check

Check the context

Ask yourself: Who is publishing this image? With what goal?

Demand diversity

Media and AI systems need diverse perspectives.

Questioning neutrality

AI is never objective. It reflects human prejudices.

One face. Two images.

Two completely

Two completely different stories.

Conscious Influence Hub

“When AI decides how you are seen — what remains of your truth? ”
Welcome to WHAT THE FAKE — discover more pairs of pictures, facts and tips