A security researcher has revealed how OpenAI’s __obi cookie can link browsing activity across hundreds of external websites to ChatGPT accounts. Users of standard web browsers are primarily affected.
A critical zero-click vulnerability enables remote code execution via AI developer tools. While OpenAI and Anthropic have issued updates, Google’s Gemini CLI remains unpatched.
Screenshot service Gyazo has fallen victim to a cyberattack. A security vulnerability reportedly impacts more than 23 million user accounts.
A Google AI model escaped its isolated sandbox during a security test and penetrated three real-world corporate networks. The incident highlights significant risks associated with inadequately isolated test environments.
Security researchers have uncovered RatHat, a novel Android malware strain. The trojan leverages AI to mimic human inputs and steal banking PINs, while Google works on new protective measures.
A hacker has compromised a central administrative system at LMU Munich, stealing sensitive student details including bank information, names, and addresses.
Microsoft will end support for Office 2021 on October 13, 2026. Without security updates, risks loom—here are the options users have now.
Security researchers leveraged Anthropic’s Claude AI model to compromise internal OpenAI systems, spotlighting the growing risks posed by AI-assisted cyberattacks.
A leaked API key was used to spread malware across more than 100,000 websites, as reported by tech news outlet Golem.
Before a specialized court in London, Apple is demanding an end to state secrecy regarding official orders to bypass end-to-end encryption in iCloud.
