Innovative techniques based on machine learning are leading advances in malware detection. A recent study introduces an active opcode insertion technique utilising reinforcement learning, allowing malware to evade detection by graph neural network (GNN) models. The system, called MalAOI, modifies malware control flow graphs while retaining functionality. Tests demonstrated a 93.73% evasion rate against GNN detection, revealing vulnerabilities in existing systems while also highlighting potential areas for future cybersecurity research and development.
Hundreds of malicious Android apps with 60 million downloads found spamming Android users with ads and stealing credentials
Bitdefender has identified over 331 potentially malicious apps on the Google Play Store that have been downloaded more than 60 million times. Part of a
