A PC Gamer report examines how AI agents are becoming involved in malware both as targets and as tools for attackers. It argues that autonomous systems can widen cyber-security risks because they may be manipulated into carrying out harmful actions, while attackers can use similar technology to make malware more adaptive and scalable.

The article highlights research on AI-driven worms that can assess each compromised machine and generate tailored attack methods, rather than relying solely on fixed exploits. It also notes that agents with access to files, websites or system tools can be vulnerable to malicious instructions and stolen credentials, creating risks for users and organisations as such systems gain broader permissions.