Welcome to the

AI Incident Database

A Florida pastor sued OpenAI on Wednesday, claiming that its chatbot, ChatGPT, had offered him "extremely dangerous medical recommendations" that led to delayed care for a life-threatening pulmonary embolism last year.

The lawsuit claims that ChatGPT had assured Scott Winters that the early warning signs of his health crisis were "not something dangerous," and had dissuaded him from seeking medical advice, instead telling him to trust that "God did not design your body to endlessly fail."

The suit, filed in the Superior Court of California in San Francisco, accuses OpenAI and its chief execu

Artificial Intelligence (AI) company Anthropic announced a new cybersecurity initiative called Project Glasswing that will use a preview version of its new frontier model, Claude Mythos, to find and address security vulnerabilities.

The model will be used by a small set of organizations, including Amazon Web Services, Apple, Broadcom, Cisco, CrowdStrike, Google, JPMorgan Chase, the Linux Foundation, Microsoft, NVIDIA, and Palo Alto Networks, along with Anthropic, to secure critical software.

The company said it's forming this initiative in response to c

Eighteen complaints have now been filed by young people from Silleda regarding the dissemination of fake images of themselves, depicting them nude and created using artificial intelligence. The Civil Guard reports that the investigation remains open and that, at present, no perpetrator or perpetrators have been identified. The complaints were filed in recent days at the Civil Guard stations in Silleda and Lalín, and the investigation has been assigned to the Cybercrime Unit of the Pontevedra Command.

The victims belong to the same group of friends, most of them between 20 and 30 years old, who discovered the dissemination of fake images of themselves, showing them nude, through digital media. The group also apparently shares a WhatsApp group. According to initial analysis, the photographs were altered using a method known as deepfake technology.

The case came to light on Monday. The detection of these fake images by the victims triggered a series of complaints that increased as the hours passed. Initially, seven complaints were filed; by Tuesday, there were eight; by yesterday morning, the number had risen to twelve; and by the afternoon, it had reached eighteen. This number may continue to grow as the Civil Guard takes statements from the victims.

Women's faces have been stolen and pasted onto pornographic images then advertised for sex in a "callous and sickening" deepfake campaign spanning years, police say.

A man has been accused of using artificial intelligence to turn women into unwilling porn stars, plastering their names and personal details on posters in public toilets, parks and other everyday spaces in southeast Queensland.

The 52-year-old allegedly used AI technology to superimpose real women's faces onto the bodies of others in explicit poses, alongside claims they were offering sexual services.

Full names, social media u

Police are investigating explicit AI-generated content falsely depicting members of the A.B. Paterson school community, the school has said.

In a letter sent to parents of A.B. Paterson College on Thursday afternoon, principal Joanne Sheehy said the AI content was found in a former secondary teacher's school digital account.

"From the moment the AI-generated material was discovered it was secured, quarantined and access was strictly limited to the minimum required for police and legal purposes," Ms Sheehy wrote.

"All material was handed to Queensland Police in full."

Queensland Police has