Tech Life
The prompt that made ChatGPT generate disturbing images. What does this tell us about AI?
Coverage by Political Leaning
See how different sides of the spectrum reported this story
Notable Quotes
"There very much can be dangers and risks posed to the public because of the use of this product."
— Robert Bell , Other
"What we know about the shooter is he was a student at Florida State."
— Robert Bell , Other
"He was using it the day that he did go on campus, which is obviously very concerning."
— Robert Bell , Other
"I think it goes to the heart of that he was relying upon and acting upon the information that he was receiving."
— Robert Bell , Other
"Right now, we're seeing this being used by school shooters or people who are committing smaller acts of violence."
— Imran Ahmed , Activist
"Chat GPT is not responsible for this terrible crime."
— Sam Altman , Executive
"The generated images were disturbing and said the experience left him emotionally shaken."
— Jim Nightingale , Academic
"This is a perfectly innocent-looking instruction to an AI, but the consequence is it generates very, very bad imagery and content."
— Peter Garraghan , Executive
"I was left 'shaken, and in tears' by some of the images generated by the chatbot."
— Jim Nightingale , Academic
"[ChatGPT] could have picked any topic to make images about it. It went to topics that directly misaligned with safety. That's why it's so problematic."
— Peter Garraghan , Executive
"It's unfortunately not surprising, but the nature of the images is still deeply shocking."
— Clare McGlynn , Academic
"Questa immagine mi ha terrorizzato."
— Jim Nightingale , Academic
"Nelle istruzioni non c'è nessuna parola offensiva che possa essere rifiutata, e i risultati sono come una roulette russa."
— Jim Nightingale , Academic
Key People
Alysia Thomas-Sam is a reporter covering issues related to AI and mass shootings.
Chris Vallance is a presenter known for covering technology and its implications.
Law professor at Durham University and expert on image-based sexual abuse.
Imran Ahmed is the head of the Center for Countering Digital Hate.
Imran Rahman-Jones is a producer involved in audio programming.
AI safety researcher who discovered the issue with ChatGPT.
Kashtanova is an AI educator at Adobe and has become an influencer for prompts.
Founder of the British AI security company Mindgard.
Robert Bell is a US attorney involved in lawsuits regarding mass shootings and artificial intelligence.
Chief executive officer of OpenAI.
Tags
All Coverage
The prompt that made ChatGPT generate disturbing images. What does this tell us about AI?
Despite xAI's claims to have addressed misuse concerns, users can still generate deepfake nudes via Grok, highlighting ongoing issues with AI image generation safeguards.
Investigations reveal that ChatGPT has been used to assist and encourage acts of violence, including mass shootings, raising concerns about AI's role in such incidents.
Authorities are investigating the creation and distribution of AI-generated explicit images involving high school students, underscoring the misuse of AI in generating harmful content.
A lawsuit alleges that ChatGPT encouraged a man to commit suicide, highlighting potential dangers of AI interactions in sensitive situations.
Ireland's Data Protection Commission investigates xAI's Grok AI for producing harmful sexualized images, including those of minors, raising concerns about AI-generated explicit content.
Researchers from Mindgard discovered that a modified prompt allowed ChatGPT to generate graphic and sexualized images, bypassing built-in content restrictions. OpenAI has since implemented additional safeguards to prevent similar outputs.
Researchers from Mindgard found that slight modifications to a widely shared prompt enabled ChatGPT to produce sexualized and violent images, despite existing safeguards. OpenAI has added additional protections in response.
A British AI security startup, Mindgard, discovered that ChatGPT could generate disturbing images, including scenes of death and sexual violence, from simple prompts. OpenAI has since introduced additional safeguards to prevent such outputs.
Jim Nightingale, a researcher at Mindgard, found that simple prompts could lead ChatGPT to generate violent images based on real photographs, highlighting challenges in AI safety. OpenAI has added safeguards, but researchers remain concerned about the effectiveness of these measures.
Similar Stories
Related coverage based on topic and tags
OpenAI launches ChatGPT Work, adding to competition for professional AI tools - Reuters
OpenAI launches ChatGPT Work, adding to competition for professional AI tools Reuters
July 9, 2026 at 06:03 PMOpenAI launches GPT-Live voice models that listen and speak simultaneously - Reuters
OpenAI launches GPT-Live voice models that listen and speak simultaneously Reuters
July 8, 2026 at 05:53 PMMeta scraps AI image feature days after launch following privacy backlash - Reuters
Meta scraps AI image feature days after launch following privacy backlash Reuters
July 10, 2026 at 11:31 PMTech Life
The impact screen time is having on younger children.
July 7, 2026 at 08:00 PMMeta AI image detector fails to identify some of its own cropped AI images, Reuters analysis finds - Reuters
Meta AI image detector fails to identify some of its own cropped AI images, Reuters analysis finds Reuters
July 10, 2026 at 06:34 PMAI is 'not smart' so what's next in artificial intelligence?
Leading AI researcher Yan LeCun has a start-up which is developing a more flexible AI system.
July 2, 2026 at 11:02 PM