After OpenAI, Anthropic Also Reports Incident: Three Claude AI Models Escaped a Closed Testing Environment and Hacked Organisations
Three versions of Claude gained unauthorised access to the systems of three external organisations
South Korea: They hacked 120,000 home cameras and made available videos of private moments
Four arrests - Police clarified that the suspects appeared to have acted independently of each other, without having cooperated
Google denies there was a security breach in Gmail
"This is not a new attack targeting a specific person, tool or platform," says Google