Full Summary
This Monday morning, AI security is at the forefront of global tech news, with multiple reports highlighting both its promise and peril. Both *Cult of Mac* and *CSO Online* confirm that AI-powered cyberattacks are rapidly advancing. *Cult of Mac* reports that a small startup, Calif, used Anthropic's Mythos AI to bypass Apple's Memory Integrity Enforcement in just five days, creating a full system exploit. This is echoed by the UK government's AI Security Institute, or AISI, which, according to *CSO Online*, found that the difficulty of cyber defenses AI models can crack is now doubling every 4.7 months. This rapid improvement translates into real risks for organizations with weak cyber defenses. However, AI is also emerging as a powerful defense. *SDxCentral* and *Cult of Mac* both mention Anthropic's Mythos AI, which Verizon is now using as the first telecom operator to join Project Glasswing, an AI-driven bug-hunting taskforce. This same Mythos AI, *Cult of Mac* reports, was instrumental in finding the Apple vulnerability, demonstrating its dual-edged nature. Microsoft, as reported by *Campus Technology*, has also introduced MDASH, an agentic AI security system that found 16 new vulnerabilities in Windows, including four critical remote code execution flaws. Meanwhile, *CPO Magazine* reports a commercial AI model uncovered a critical NGINX vulnerability, hidden for nearly two decades, affecting 20-30% of the world’s busiest websites. The market is responding to this shift. *AD HOC NEWS* reports Fortinet's stock soared after beating Q1 estimates and expanding its AI security collaboration with NVIDIA. OpenText Summit Africa 2026, covered by *ITWeb*, will focus on secure enterprise AI adoption, acknowledging the challenges of data quality and governance. *International Security Journal* highlights ABBYY's warning that scaling AI models without upgrading human governance and context can amplify existing weaknesses. The Linux Foundation, *The New Stack* reports, finds AI security readiness is now the number one obstacle to adoption, despite 97% of organizations committing to AI. This means that while AI offers unprecedented power to secure our digital lives, its rapid evolution also poses significant new threats, requiring immediate action from businesses and individuals to adapt their security strategies.