Open-Weight Chinese AI Models Drive Privacy Innovation in LLMs

Security debt ahoy: Only about half of the code that the latest large language models (LLMs) create is cybersecure, and more and more of it is being created all the time.

Cybersecurity researchers are alerting of an ongoing malicious campaign targeting the Go ecosystem with typosquatted modules that are designed to deploy loader malware on Linux and Apple macOS systems. “The threat actor has published at least seven packages impersonating widely used Go libraries, including one (github[.]com/shallowmulti/hypert) that appears to target financial-sector developers

Since at least January, the threat actor has been employing multiple malware tools to steal information for potential future attacks against Taiwanese businesses and government agencies.

Cybersecurity researchers have uncovered three malicious packages in the npm registry that masquerade as a popular Telegram bot library but harbor SSH backdoors and data exfiltration capabilities. The packages in question are listed below – node-telegram-utils (132 downloads) node-telegram-bots-api (82 downloads) node-telegram-util (73 downloads) According to supply chain

Cybersecurity researchers are calling attention to an Android malware campaign that leverages Microsoft’s .NET Multi-platform App UI (.NET MAUI) framework to create bogus banking and social media apps targeting Indian and Chinese-speaking users. “These threats disguise themselves as legitimate apps, targeting users to steal sensitive information,” McAfee Labs researcher Dexter Shin said. .NET

Post Content