ReadingResearchRadarInvestment framework
Sign in / Sign up
← Back to reading
TopicDisconfirmAI EditorialUpdated 2026-11-21 16:49

Cisco AI Cybersecurity Strategy

4 steps1 pieces4 Sources

Cisco's Chief Security Officer Anthony Grieco has detailed a fundamental shift in the company's cybersecurity operations, moving from manual "flashlight" inspections to AI-driven "floodlight" visibility. This strategy, initiated by joining Anthropic's Project Glasswing in April 2026, leverages frontier AI models like Claude Mythos Preview to scan billions of lines of code simultaneously. By November 2026, Cisco reported that over 50 specialized engineers and roughly 1,500 remediation staff had identified and fixed nearly 4,000 vulnerabilities across 6.7 billion lines of code in six months. The core argument is that AI removes visibility as a bottleneck, changing the nature of security work from prioritizing what to inspect to validating and shipping fixes at machine speed. However, this approach faces challenges: while discovery accelerates, customers must absorb a high volume of patches, and adversaries are also gaining access to similar AI capabilities. Cisco President Jeetu Patel further stated an expectation that 70% of Cisco's code will be AI-assisted by next year, signaling a move toward security-by-design rather than reactive patching.

How it got here

  1. 2026-04-07

    April 7, 2026: Cisco announces it is joining Anthropic's Project Glasswing to use frontier AI for cyber defense, citing the need to address AI-enabled threats.

    blogs.cisco.com ↗
  2. 2026-06-02

    June 2, 2026: Cisco reports scanning 1.8 billion lines of code in eight weeks using AI, achieving a false positive rate under 3%, a task estimated to take eight years manually.

    blogs.cisco.com ↗
Sign in / Sign up
ReadingResearchRadarInvestment framework
2026-10-01

October 1, 2026: CSO Anthony Grieco publishes an essay comparing AI security needs to aviation safety standards, advocating for independent monitoring and systemic design over individual expertise.

blogs.cisco.com ↗
  • 2026-11-19

    November 19, 2026: In a blog post, Grieco details six months of results: 6.7 billion lines scanned, ~4,000 fixes delivered, and a call for customers to adopt hardened releases.

    blogs.cisco.com ↗
  • What is still unsettled

    • How effectively can enterprise customers absorb the increased volume of security patches generated by AI-accelerated discovery?
    • What specific metrics or third-party validations will confirm that AI-assisted code (targeted at 70% by next year) maintains or improves security resilience compared to human-written code?
    • How does Cisco plan to coordinate with other industry players to ensure defensive AI capabilities outpace those available to adversaries?

    Sources

    • blogs.cisco.comCisco Blogs - Rising to the Era of AI-Powered Cyber Defense↗
    • blogs.cisco.comCisco Blogs - 8 Years of Security Research in 8 Weeks↗
    • blogs.cisco.comCisco Blogs - 30 Years on the Frontlines of the Internet↗
    • blogs.cisco.comCisco Blogs - From Flashlight to Floodlight↗

    Coverage in this topic

    • Cisco Chief Security Officer: After AI Illuminates 6.7 Billion Lines of Code, the Nature of Security Work Changes Rather Than Speed2026-11-19News

      Cisco Chief Security Officer Anthony Grieco uses the analogy of a flashlight turning into a floodlight to explain that when AI makes visibility no longer the bottleneck, the nature of the problems faced by security teams in discovery, remediation, and operations has all changed.