Data Privacy
Study uncovers 58 billion data leaks since 2004
由 AI 报道
A new report highlights the massive scale of online data breaches over the past two decades. Nearly 58 billion personal data points have been exposed since 2004. The findings underscore ongoing privacy risks in the digital age.
Claude AI vulnerable to prompt injection data theft
Security researchers have found that Anthropic's Claude AI can be manipulated through prompt injection to send private company data to hackers. The attack requires only persuasive language to trick the model. This vulnerability highlights risks in AI systems handling sensitive information.
Pixnapping attack steals sensitive data from Android devices
由 AI 报道
Academic researchers have revealed a new Android vulnerability called Pixnapping that lets malicious apps steal 2FA codes, private messages, and other visible data without needing permissions. The attack, which takes less than 30 seconds, exploits screen rendering times and has been demonstrated on Google Pixel phones and the Samsung Galaxy S25. Google has released partial mitigations, with further patches planned.