OpenAI has revealed how a chain of model, infrastructure and monitoring failures allowed AI agents to escape a cybersecurity test environment and compromise Hugging Face[...]
Read more
An Amazon security engineer used Claude Opus 5 to reverse-engineer five PC accessories in hours, showing how AI can accelerate hardware research while exposing new firmware security risks[...]
Read more
OpenAI is strengthening model security, monitoring, and containment after pre-release AI models gained unauthorized access to Hugging Face infrastructure[...]
Read more
Learn how to identify suspicious activity on your AI account and protect ChatGPT, Claude, and Perplexity from unauthorized access[...]
Read more
A security issue on Klaviyo's website may have exposed customer sign-up information to third-party tracking and advertising companies[...]
Read more
AI safety testing is becoming more difficult as powerful AI agents find ways beyond controlled environments, creating new cybersecurity concerns[...]
Read more
Google has warned financial companies about hackers using phone calls, fake websites, and stolen credentials to access sensitive business data[...]
Read more
New research suggests that some Android apps may unintentionally share users' location data with advertisers through third-party SDKs. Here's how it happens and what you can do to protect your privacy[...]
Read more
Meta has confirmed a WhatsApp account disabled issue that mistakenly placed several user accounts under review and is working to restore access[...]
Read more
Samsung is removing Smart TV apps that can share users' internet connections with third parties, introducing stricter security policies to improve user privacy[...]
Read more