AI-powered malware emerges as significant threat, AV-TEST reports

As Cybersecurity Awareness Month 2026 approaches, cybersecurity firm AV-TEST highlights a significant shift in the threat landscape: the growing use of artificial intelligence (AI) by cybercriminals to develop and deploy malware and conduct cyberattacks.

AI is accelerating and scaling known attack methods, making them cheaper and more accessible to less technically skilled threat actors, AV-TEST reports. Erik Heyland, Head of Testing Labs at AV-TEST, noted that nearly 75 percent of newly discovered malware samples exhibit AI-driven polymorphism. The company anticipates close to 100 million new malware samples for Windows in 2026 alone, with its databases already exceeding one billion.

Industry reports from companies like CrowdStrike and Check Point indicate a substantial rise in AI-powered and agentic AI attacks. AI's role is evolving from an assistant to an active operator in cyberattacks. Analysis shows that a significant portion of malicious accounts, around 67.3 percent, leveraged AI for writing malicious code or preparing attacks.

AI is also impacting vulnerability discovery and exploit code creation. The International AI Safety Report detailed a competition where an AI agent identified 77 percent of software vulnerabilities and generated exploit code for them.

In response, security vendors are integrating AI-enhanced modules into their detection systems. AV-TEST's regular evaluations demonstrate that current security products are largely capable of defending against these AI-driven threats, though the arms race between attackers and defenders continues as AI technology rapidly evolves.