Researchers uncover malware that uses AI to choose its next move

Help Net Security reports multiple cybersecurity developments, including active exploitation of CVE-2026-7273 against nearly 1,000 Zyxel GS1900 switches across 48 countries, domain hijacking linked to a LAPSUS$-branded page, AI-enabled malware research, deepfake social-engineering incidents, fraudulent AI subscription sites using Google sign-in, and research into 5G subscriber tracking via fake base stations. The most urgent item is the widespread exploitation of the Zyxel vulnerability and associated data exfiltration.