Cisco AI Security Push: Enterprise Trust Gap Amidst New Threats

Cisco officials are urging customers to embrace new agentic defenses as a new era in cybersecurity emerges. This comes as frontier AI models, designed to find vulnerabilities, are set to be released. One such model, Anthropic Claude Mythos, was first revealed in April. Its initial preview was limited to about 50 major organizations. During this preview, partners found over 10,000 high- or critical-severity security flaws in common applications. This included more than 1,000 open-source projects. Project Glasswing has now expanded to 150 partners, and Mythos is nearing general availability. This means attackers will soon be able to use similar tools to find vulnerabilities. Cisco executive Liz Centoni warns that AI tools are already mapping networks at machine speed. Cisco IQ, a new SaaS tool, aims to help. It provides an automated, real-time asset inventory to find vulnerable network devices. This tool has been used by over 2,000 customers since late April. It also streamlines support requests. In July, Cisco IQ will add on-premises support and other features. The bottom line is that advanced AI tools are changing the landscape of cybersecurity for everyone.