German police¹ is using ChatGPT to fake images of injured police officers.
-
Imho this is not about what position we hold on football hooligans or on violence against police under different circumstances.
This is about whether the police can be trusted (ha!) if it uses AI fakes to “illustrate” acts of violence. There might even be (legal, moral or other) reasons for not using actual footage - but then just report on the facts without a photo. It’s that simple.
@alineblankertz Yes! Better no picture than an AI-generated one.
While you were writing your comment, I completely revamped my initial post. I don't know if you already saw it. I had to admit that I wrote my initial post too rashly.
-
German police¹ is using ChatGPT to fake images of injured police officers.
After a soccer match with believable hooligan violence against police officers, the German union of police officers (GdP) rightly condemns violence against police officers.
However, instead of using real footage for their press release² of the actual violence that was actually committed, they used some AI-generated fake image and denoted that fact only with a teeny-tiny note in the corner, saying "AI-generated: ChatGPT".
I find this very problematic.
The news report mentioned several injured police officers, some badly injured, and the fake image shows an officer bleeding from his forehead. I have no way to know if any officer was actually bleeding from his head or if this is an exaggeration to make a point. In a time when politicians call for ever more surveillance and ever more power for police officers, this can function as dangerous misinformation.
Especially considering that the Nazis are growing ever stronger and our so-called democratic political parties are building a police state for them. Manufacturing consent using fake images of violence against police officers is dangerous.
Although this was a soccer match, not an anti-genocide protest, and soccer matches occasionally attract violent hooligans, protests are almost always peaceful. At least until the police starts the violence and beats up anyone who protests against fossil fuels or genocide. If we allow police to use AI-generated images in their press releases, then we cannot believe them anything anymore at all. If people still believe them, anyway.
Condemning violence only when it hits them is also hypocritical.
Police is already instigating and inciting against imagined "violent radical leftists", even before/without the use of AI fakes.
The GdP isn't innocent. Just two months ago when the modern NSDAP wanted to start their Hitler Youth equivalent and people peacefully blocked the roads, whereafter the police played taxi for officially recognized Nazis, the GdP said "the radical leftists showed their ugly face and violence".
It's inacceptable for the police to fake images of violence against police.
¹ Sachsen, not even the infamous Berlin police
² https://www.gdp.de/sachsen/de/stories/2026/01/gewalt-im-fussballspiel-magdeburg-gegen-dresden@davidculley Ich stimme dem zu, dass zur Darstellung der erlittenen Gewalt der #Polizei durch die Hooligans keine KI-Bilder von der Polizei oder anderen staatlichen Stellen verbreitet werden sollten. Niemand braucht sowas! Habe sowas hier auch schon von einer politischen Partei gesehen, die reale Politiker mit KI verfälschte. Ich bin grundsätzlich gegen "Fotos" und Videos von #KI. Diese Art der Verfälschung als angebliches Foto zerstört das Vertrauen in Fotos allgemein.
Heute glauben wir noch den Handy-Videos aus Minneanapolis. Doch werden wir das noch in 10 Jahren können? KI zerstört hiermit die #Demokratie. -
@davidculley Why the “it’s just symbolic / AI” defense rings hollow
Because the harm isn’t about deception in a narrow sense (“this exact event happened”). It’s about legitimizing a power narrative:
We are the injured, reasonable authority; they are the threatening crowd.
AI just makes it cheaper, faster, and safer for institutions to manufacture that affect without accountability.
I was able to generate almost the exact same image and submitted a report that ChatGPT shouldn’t be making propaganda.
They answered and indicated no policy was violated.
️ -
German police¹ is using ChatGPT to fake images of injured police officers.
After a soccer match with believable hooligan violence against police officers, the German union of police officers (GdP) rightly condemns violence against police officers.
However, instead of using real footage for their press release² of the actual violence that was actually committed, they used some AI-generated fake image and denoted that fact only with a teeny-tiny note in the corner, saying "AI-generated: ChatGPT".
I find this very problematic.
The news report mentioned several injured police officers, some badly injured, and the fake image shows an officer bleeding from his forehead. I have no way to know if any officer was actually bleeding from his head or if this is an exaggeration to make a point. In a time when politicians call for ever more surveillance and ever more power for police officers, this can function as dangerous misinformation.
Especially considering that the Nazis are growing ever stronger and our so-called democratic political parties are building a police state for them. Manufacturing consent using fake images of violence against police officers is dangerous.
Although this was a soccer match, not an anti-genocide protest, and soccer matches occasionally attract violent hooligans, protests are almost always peaceful. At least until the police starts the violence and beats up anyone who protests against fossil fuels or genocide. If we allow police to use AI-generated images in their press releases, then we cannot believe them anything anymore at all. If people still believe them, anyway.
Condemning violence only when it hits them is also hypocritical.
Police is already instigating and inciting against imagined "violent radical leftists", even before/without the use of AI fakes.
The GdP isn't innocent. Just two months ago when the modern NSDAP wanted to start their Hitler Youth equivalent and people peacefully blocked the roads, whereafter the police played taxi for officially recognized Nazis, the GdP said "the radical leftists showed their ugly face and violence".
It's inacceptable for the police to fake images of violence against police.
¹ Sachsen, not even the infamous Berlin police
² https://www.gdp.de/sachsen/de/stories/2026/01/gewalt-im-fussballspiel-magdeburg-gegen-dresden@davidculley
Any reasonable person would see their treachery as a reason to withhold sympathy instead of giving it.That is the true loss of the schemers. They shoot themselves in the foot with this.
-
@davidculley Ich stimme dem zu, dass zur Darstellung der erlittenen Gewalt der #Polizei durch die Hooligans keine KI-Bilder von der Polizei oder anderen staatlichen Stellen verbreitet werden sollten. Niemand braucht sowas! Habe sowas hier auch schon von einer politischen Partei gesehen, die reale Politiker mit KI verfälschte. Ich bin grundsätzlich gegen "Fotos" und Videos von #KI. Diese Art der Verfälschung als angebliches Foto zerstört das Vertrauen in Fotos allgemein.
Heute glauben wir noch den Handy-Videos aus Minneanapolis. Doch werden wir das noch in 10 Jahren können? KI zerstört hiermit die #Demokratie.> KI zerstört hiermit die Demokratie.
Sie zerstört noch viel mehr als nur die Demokratie: sie hat die Macht, den Glauben an alles zu zerstören, weil man schliesslich nichts mehr glauben kann. Es entsteht eine kafkaeske Welt.
-
> KI zerstört hiermit die Demokratie.
Sie zerstört noch viel mehr als nur die Demokratie: sie hat die Macht, den Glauben an alles zu zerstören, weil man schliesslich nichts mehr glauben kann. Es entsteht eine kafkaeske Welt.
-
@davidculley LLMs are the tools of fascism.
The tools of anyone who wants to fool anyone about anything: communists, Islamicists, con artists, phishers, students who want to fool their examiners, crooks of any kind.
-
The tools of anyone who wants to fool anyone about anything: communists, Islamicists, con artists, phishers, students who want to fool their examiners, crooks of any kind.
@watcher2026 You're talking nonsense.
-
@Su_G @davidculley
the image prompt and responses from the LLM are simply continuations of the tone I used, so it's merely agreeing with me, not stating policy. -
@Su_G @davidculley
the image prompt and responses from the LLM are simply continuations of the tone I used, so it's merely agreeing with me, not stating policy.@Netraven I might have been responsible for spreading misinformation. The news is about soccer hooligans, not anti-genocide protestors. I edited my initial post since first publishing it. Sorry for any confusion.
But in your image, it even knew to use BE for Berlin cops. That's damning. Many videos of infamous Berlin police on social media.
-
@Netraven I might have been responsible for spreading misinformation. The news is about soccer hooligans, not anti-genocide protestors. I edited my initial post since first publishing it. Sorry for any confusion.
But in your image, it even knew to use BE for Berlin cops. That's damning. Many videos of infamous Berlin police on social media.
@davidculley ya, either way, it shouldn't be doing that. Wanna take bets if it will allow me to do the same thing for ICE?
-
> KI zerstört hiermit die Demokratie.
Sie zerstört noch viel mehr als nur die Demokratie: sie hat die Macht, den Glauben an alles zu zerstören, weil man schliesslich nichts mehr glauben kann. Es entsteht eine kafkaeske Welt.
@watcher2026 @werawelt @davidculley
Darauf ist die Menschheit nicht vorbereitet -
@davidculley@hachyderm.io Seems like Germany and the US are lockstep marching towards a new world order....
@gme @davidculley tacuisses! You startled the butterfly.
-
No idea why they posted an AI generated photo, but there were still a lot of injured officers.
https://www.tagesschau.de/inland/regional/sachsenanhalt/ausschreitungen-magdeburg-102.html
@simonriley @wakame @davidculley Immer daran denken: Da steht nicht, wodurch die Verletzung ausgelöst wurde.
Ich nehme mit Sicherheit keine Hools in Schutz, aber schon wenn jemand sich einen Fuß umknickt oder das eigene Tränengas einatmet zählt das in die Statistik.
Eigentlich unverständlich, dass das immer noch als Benchmark für die Schwere von Krawallen genommen wird.
-
German police¹ is using ChatGPT to fake images of injured police officers.
After a soccer match with believable hooligan violence against police officers, the German union of police officers (GdP) rightly condemns violence against police officers.
However, instead of using real footage for their press release² of the actual violence that was actually committed, they used some AI-generated fake image and denoted that fact only with a teeny-tiny note in the corner, saying "AI-generated: ChatGPT".
I find this very problematic.
The news report mentioned several injured police officers, some badly injured, and the fake image shows an officer bleeding from his forehead. I have no way to know if any officer was actually bleeding from his head or if this is an exaggeration to make a point. In a time when politicians call for ever more surveillance and ever more power for police officers, this can function as dangerous misinformation.
Especially considering that the Nazis are growing ever stronger and our so-called democratic political parties are building a police state for them. Manufacturing consent using fake images of violence against police officers is dangerous.
Although this was a soccer match, not an anti-genocide protest, and soccer matches occasionally attract violent hooligans, protests are almost always peaceful. At least until the police starts the violence and beats up anyone who protests against fossil fuels or genocide. If we allow police to use AI-generated images in their press releases, then we cannot believe them anything anymore at all. If people still believe them, anyway.
Condemning violence only when it hits them is also hypocritical.
Police is already instigating and inciting against imagined "violent radical leftists", even before/without the use of AI fakes.
The GdP isn't innocent. Just two months ago when the modern NSDAP wanted to start their Hitler Youth equivalent and people peacefully blocked the roads, whereafter the police played taxi for officially recognized Nazis, the GdP said "the radical leftists showed their ugly face and violence".
It's inacceptable for the police to fake images of violence against police.
¹ Sachsen, not even the infamous Berlin police
² https://www.gdp.de/sachsen/de/stories/2026/01/gewalt-im-fussballspiel-magdeburg-gegen-dresden@davidculley Well, should they have used German based #Flux?
-
German police¹ is using ChatGPT to fake images of injured police officers.
After a soccer match with believable hooligan violence against police officers, the German union of police officers (GdP) rightly condemns violence against police officers.
However, instead of using real footage for their press release² of the actual violence that was actually committed, they used some AI-generated fake image and denoted that fact only with a teeny-tiny note in the corner, saying "AI-generated: ChatGPT".
I find this very problematic.
The news report mentioned several injured police officers, some badly injured, and the fake image shows an officer bleeding from his forehead. I have no way to know if any officer was actually bleeding from his head or if this is an exaggeration to make a point. In a time when politicians call for ever more surveillance and ever more power for police officers, this can function as dangerous misinformation.
Especially considering that the Nazis are growing ever stronger and our so-called democratic political parties are building a police state for them. Manufacturing consent using fake images of violence against police officers is dangerous.
Although this was a soccer match, not an anti-genocide protest, and soccer matches occasionally attract violent hooligans, protests are almost always peaceful. At least until the police starts the violence and beats up anyone who protests against fossil fuels or genocide. If we allow police to use AI-generated images in their press releases, then we cannot believe them anything anymore at all. If people still believe them, anyway.
Condemning violence only when it hits them is also hypocritical.
Police is already instigating and inciting against imagined "violent radical leftists", even before/without the use of AI fakes.
The GdP isn't innocent. Just two months ago when the modern NSDAP wanted to start their Hitler Youth equivalent and people peacefully blocked the roads, whereafter the police played taxi for officially recognized Nazis, the GdP said "the radical leftists showed their ugly face and violence".
It's inacceptable for the police to fake images of violence against police.
¹ Sachsen, not even the infamous Berlin police
² https://www.gdp.de/sachsen/de/stories/2026/01/gewalt-im-fussballspiel-magdeburg-gegen-dresden@davidculley
Not to forget, the injured policemen statistics include the ones who have sprained their ankle when exiting their vehicle or tripped over stuff or just feel down without anyone's involvement. -
@simonriley @wakame @davidculley Immer daran denken: Da steht nicht, wodurch die Verletzung ausgelöst wurde.
Ich nehme mit Sicherheit keine Hools in Schutz, aber schon wenn jemand sich einen Fuß umknickt oder das eigene Tränengas einatmet zählt das in die Statistik.
Eigentlich unverständlich, dass das immer noch als Benchmark für die Schwere von Krawallen genommen wird.
@eingemaischt @wakame @davidculley
Berechtigter Einwand, wir wissen nicht, was die Verletzungen genau waren.
Dass aber BeamtInnen zur Behandlung ins Krankenhaus gehen, ist dennoch eher ungewöhnlich. Die von dir angesprochenen Verletzungen werden, wenn überhaupt, ambulant behandelt.
-
@eingemaischt @wakame @davidculley
Berechtigter Einwand, wir wissen nicht, was die Verletzungen genau waren.
Dass aber BeamtInnen zur Behandlung ins Krankenhaus gehen, ist dennoch eher ungewöhnlich. Die von dir angesprochenen Verletzungen werden, wenn überhaupt, ambulant behandelt.
@simonriley @wakame @davidculley das sind dann aber auch nur 16 statt 64 aus der Headline. Und selbst die Zahl ist manipulierbar.
Noch einmal det Disclaimer : Ich weiß nicht, was hier abging. Es geht mir nur darum zu betonen, dass die Zahl der Verletzten kein Benchmark ist.
-
@davidculley@hachyderm.io
That's gotta be against the GDPR or some other European legislation about government propaganda or some shit.@gme @davidculley This might to be covered by EU Regulation 2024/1689 "KI-Verordnung" somewhere in Article 5.
